Release 1.7.0: Grafana catalog, ingress/IRM, tests

- Dockerfile + entrypoint (alembic + uvicorn), compose с healthcheck
- .gitea/workflows: ci (pytest), deploy (SSH + compose по тегу v*)
- docs/CICD.md: секреты, pvestandt9, ручной откат через workflow_dispatch

Made-with: Cursor

.dockerignore

@@ -0,0 +1,13 @@
+.git
+.github
+.gitea
+.venv
+__pycache__
+*.pyc
+.pytest_cache
+.env
+.env.*
+web/node_modules
+web/dist
+*.md
+!README.md

.env.example

@@ -3,10 +3,15 @@
 HTTP_ADDR=0.0.0.0:8080
 LOG_LEVEL=info
 
-# Опционально: если задан — POST /api/v1/ingress/grafana требует заголовок X-OnGuard-Secret
+# Опционально: общий секрет для вебхуков (если у источника в JSON не задан свой webhook_secret)
 # GRAFANA_WEBHOOK_SECRET=
 
-# --- Grafana HTTP API (service account, не пароль admin) ---
+# Несколько Grafana: JSON-массив. slug — часть URL вебхука: /api/v1/ingress/grafana/{slug}
+# Пример: [{"slug":"adibrov","api_url":"https://grafana-adibrov.example","api_token":"glsa_...","webhook_secret":"длинный-секрет"}]
+# Если пусто, но задан GRAFANA_URL — один источник со slug "default" (вебхук /api/v1/ingress/grafana/default)
+# GRAFANA_SOURCES_JSON=
+
+# --- Grafana HTTP API (один инстанс, если без GRAFANA_SOURCES_JSON) ---
 # URL без завершающего слэша. Токен: Grafana → Administration → Service accounts → onguard24 → Add service account token
 GRAFANA_URL=https://grafana.pvenode.ru
 # GRAFANA_SERVICE_ACCOUNT_TOKEN=

.gitea/workflows/ci.yml

@@ -0,0 +1,24 @@
+# Forgejo / Gitea Actions — проверка перед деплоем
+name: CI
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Python 3.12
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Pytest
+        run: |
+          pip install -e ".[dev]"
+          pytest -q

.gitea/workflows/deploy.yml

@@ -0,0 +1,48 @@
+# Деплой на сервер по SSH после пуша тега v* или вручную (в т.ч. откат на старый тег).
+name: Deploy
+
+on:
+  push:
+    tags:
+      - "v*"
+  workflow_dispatch:
+    inputs:
+      ref:
+        description: "Git ref (тег для релиза или отката, напр. v1.5.0 или v1.4.1)"
+        required: true
+        default: "main"
+
+jobs:
+  deploy:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Определить ревизию
+        id: pick
+        run: |
+          if [ "${{ github.event_name }}" = "workflow_dispatch" ]; then
+            echo "ref=${{ inputs.ref }}" >> "$GITHUB_OUTPUT"
+          else
+            echo "ref=${{ github.ref_name }}" >> "$GITHUB_OUTPUT"
+          fi
+
+      - name: SSH — fetch, checkout, docker compose
+        uses: appleboy/ssh-action@v1.2.0
+        with:
+          host: ${{ secrets.DEPLOY_HOST }}
+          username: ${{ secrets.DEPLOY_USER }}
+          key: ${{ secrets.DEPLOY_SSH_KEY }}
+          script_stop: true
+          command_timeout: 20m
+          script: |
+            set -euo pipefail
+            cd "${{ secrets.DEPLOY_PATH }}"
+            git fetch origin --tags --prune
+            git checkout "${{ steps.pick.outputs.ref }}"
+            if git show-ref --verify --quiet "refs/remotes/origin/${{ steps.pick.outputs.ref }}"; then
+              git reset --hard "origin/${{ steps.pick.outputs.ref }}"
+            else
+              git reset --hard "${{ steps.pick.outputs.ref }}"
+            fi
+            docker compose build
+            docker compose up -d
+            docker compose ps

CHANGELOG.md

@@ -2,6 +2,47 @@
 
 Формат: семантическое версионирование `MAJOR.MINOR.PATCH`. Git-теги `v1.0.0`, `v1.1.0` и т.д. — см. [docs/VERSIONING.md](docs/VERSIONING.md).
 
+## [1.7.0] — 2026-04-03
+
+Каталог Grafana (топология правил), доработки ingress/IRM, тесты.
+
+### Добавлено
+
+- **Alembic `003_ingress_org_service`**, **`004_grafana_catalog`** — метаданные и снимок папок/правил Grafana.
+- **Модуль `grafana-catalog`:** `POST …/sync`, `GET …/meta`, `GET …/tree`, UI.
+- **`onguard24/integrations/grafana_topology.py`**, **`grafana_sources.py`**.
+- **Документация:** [docs/GRAFANA_TOPOLOGY.md](docs/GRAFANA_TOPOLOGY.md).
+- **Тесты:** `test_grafana_topology.py`, `test_grafana_catalog_api.py`, `irm_db_fake.py`, `test_irm_api_with_fake_db.py`.
+
+## [1.6.0] — 2026-04-03
+
+Docker-образ, `docker-compose.yml`, CI/CD Forgejo/Gitea Actions.
+
+### Добавлено
+
+- **`Dockerfile`**, **`docker-compose.yml`**, **`deploy/entrypoint.sh`** — `alembic upgrade` + `uvicorn` (отключение: `SKIP_ALEMBIC=1`).
+- **`.gitea/workflows/ci.yml`** — pytest на push в `main` и PR.
+- **`.gitea/workflows/deploy.yml`** — деплой по пушу тега `v*` или вручную; **откат** = тот же workflow с `ref` = старый тег.
+- **[docs/CICD.md](docs/CICD.md)** — секреты, подготовка `root@pvestandt9`, порядок релиза и отката.
+
+## [1.5.0] — 2026-04-03
+
+IRM-ядро: инциденты, задачи, эскалации, миграция БД, документация.
+
+### Добавлено
+
+- **Документация:** [docs/IRM.md](docs/IRM.md) — матрица функций IRM и что настраивать в Grafana.
+- **Alembic `002_irm_core`:** таблицы `incidents`, `tasks`, `escalation_policies`.
+- **Модули:** `incidents` (API + UI, авто-создание из `alert.received` при наличии БД), `tasks`, `escalations`.
+- **`register_module_events(bus, pool)`** — подписки получают пул PostgreSQL.
+- **Тесты:** `tests/test_irm_modules.py`, обновлены тесты навигации.
+
+## [1.4.1] — 2026-04-03
+
+### Исправлено
+
+- **Alembic:** к URL с `postgresql+psycopg` добавляется `client_encoding=utf8`, чтобы `alembic upgrade head` не падал с `TypeError: ... bytes-like object` при некоторых настройках кодировки на стороне PostgreSQL/psycopg3.
+
 ## [1.4.0] — 2026-04-03
 
 Правое меню «Разделы» на главной и на страницах модулей, пункты из `MODULE_MOUNTS` (`title` + `ui_router`).

Dockerfile

@@ -0,0 +1,20 @@
+FROM python:3.12-slim-bookworm
+
+WORKDIR /app
+
+RUN apt-get update \
+    && apt-get install -y --no-install-recommends libpq5 \
+    && rm -rf /var/lib/apt/lists/*
+
+COPY pyproject.toml README.md alembic.ini ./
+COPY alembic ./alembic
+COPY onguard24 ./onguard24
+COPY deploy/entrypoint.sh /entrypoint.sh
+RUN chmod +x /entrypoint.sh \
+    && pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir .
+
+ENV PYTHONUNBUFFERED=1
+EXPOSE 8080
+
+ENTRYPOINT ["/entrypoint.sh"]

README.md

@@ -1,6 +1,6 @@
 # onGuard24
 
-**Версия: 1.4.0** · Модульный монолит на **Python (FastAPI)**: ядро, приём алертов из Grafana, заготовки модулей (дежурства, контакты, светофор), PostgreSQL, проверки Vault / Grafana / Forgejo.
+**Версия: 1.6.0** · Модульный монолит на **Python (FastAPI)**: ядро, приём алертов из Grafana, заготовки модулей (дежурства, контакты, светофор), PostgreSQL, проверки Vault / Grafana / Forgejo.
 
 | Документ | Назначение |
 |----------|------------|
@@ -10,6 +10,8 @@
 | [docs/AI_CONTEXT.md](docs/AI_CONTEXT.md) | Краткий контекст для доработок |
 | [docs/DOMAIN.md](docs/DOMAIN.md) | Сущности (инцидент, алерт, эскалация), шина событий |
 | [docs/MODULES.md](docs/MODULES.md) | Как добавлять модули и подписки на события |
+| [docs/IRM.md](docs/IRM.md) | Функционал IRM: что делаем, что в Grafana |
+| [docs/CICD.md](docs/CICD.md) | Forgejo Actions: деплой на сервер, откат по тегу |
 
 **Репозиторий:** [forgejo.pvenode.ru/admin/onGuard24](https://forgejo.pvenode.ru/admin/onGuard24)
 
@@ -21,7 +23,7 @@
 - **PostgreSQL:** пул asyncpg, таблица `ingress_events` для сырых тел webhook Grafana; схема через **Alembic** (отдельные ревизии в `alembic/versions/`).
 - **POST `/api/v1/ingress/grafana`** — приём JSON алерта (опционально защита `X-OnGuard-Secret`).
 - **GET `/`**, **GET `/api/v1/status`** — проверки: БД, Vault, Grafana (service account), Forgejo (PAT).
-- **Модули (API + веб-UI):** JSON под `/api/v1/modules/...`, полные HTML-страницы под `/ui/modules/<slug>/`, превью на главной `/` (см. [docs/MODULES.md](docs/MODULES.md)).
+- **Модули (API + веб-UI):** IRM — **инциденты**, **задачи**, **эскалации**, дежурства, контакты, светофор; JSON под `/api/v1/modules/...`, UI под `/ui/modules/<slug>/` (см. [docs/MODULES.md](docs/MODULES.md), [docs/IRM.md](docs/IRM.md)).
 - **Фронт (опционально):** `web/` — Vite + React, прокси на API.
 
 Чего **ещё нет** (следующие версии): авторизация публичных API (кроме секрета webhook), полноценная бизнес-логика IRM в коде (эскалации, дежурства, светофор), фоновые задачи. Доменные сущности и задел под модули описаны в [docs/DOMAIN.md](docs/DOMAIN.md).
@@ -35,7 +37,7 @@ source .venv/bin/activate   # Windows: .venv\Scripts\activate
 pip install -e .
 ```
 
-Скопируй `.env.example` в `.env` и заполни секреты (см. ниже). Перед первым запуском с БД примените миграции:
+Скопируй `.env.example` в `.env` и заполни секреты (см. ниже). Перед первым запуском с БД примените миграции (в т.ч. таблицы IRM: `incidents`, `tasks`, `escalation_policies`):
 
 ```bash
 alembic upgrade head
@@ -82,7 +84,7 @@ Vite проксирует `/api` на `http://127.0.0.1:8080` (см. `web/vite.c
 
 ## Миграции БД (Alembic)
 
-- URL БД: переменная **`DATABASE_URL`** (как у приложения; в `alembic/env.py` используется синхронный драйвер `postgresql+psycopg`).
+- URL БД: переменная **`DATABASE_URL`** (как у приложения; в `alembic/env.py` используется синхронный драйвер `postgresql+psycopg`). К URL автоматически добавляется **`client_encoding=utf8`**, чтобы миграции не падали с `TypeError: ... bytes-like object`, если на сервере PostgreSQL включён режим вроде `SQL_ASCII`.
 - Применить схему: `alembic upgrade head`.
 - Новая ревизия: `alembic revision -m "описание"` и правка файла в `alembic/versions/`.
 
@@ -96,3 +98,11 @@ pytest
 ```
 
 Покрытие: `/health`, `/api/v1/status`, webhook Grafana; внешние вызовы (Vault, Grafana, Forgejo) в тестах статуса подменяются моками.
+
+## Docker и CI/CD
+
+```bash
+docker compose build && docker compose up -d
+```
+
+Деплой через **Forgejo Actions** (тег `v*`, SSH на сервер): см. [docs/CICD.md](docs/CICD.md). **Откат:** вручную запустить workflow **Deploy** с полем `ref` = нужный старый тег.

alembic/env.py

@@ -20,6 +20,14 @@ if config.config_file_name is not None:
 target_metadata = None
 
 
+def _ensure_psycopg_client_encoding(url: str) -> str:
+    """Иначе psycopg3 при SQL_ASCII на сервере отдаёт version() как bytes → падает SQLAlchemy re.match."""
+    if "+psycopg" not in url or "client_encoding=" in url:
+        return url
+    join = "&" if "?" in url else "?"
+    return f"{url}{join}client_encoding=utf8"
+
+
 def get_sync_url() -> str:
     url = os.environ.get("DATABASE_URL", "").strip()
     if not url:
@@ -30,7 +38,7 @@ def get_sync_url() -> str:
         url = url.replace("postgresql://", "postgresql+psycopg://", 1)
     if "+asyncpg" in url:
         url = url.replace("+asyncpg", "+psycopg")
-    return url
+    return _ensure_psycopg_client_encoding(url)
 
 
 def run_migrations_offline() -> None:

alembic/versions/002_irm_core_tables.py

@@ -0,0 +1,72 @@
+"""irm core: incidents, tasks, escalation_policies
+
+Revision ID: 002_irm_core
+Revises: 001_initial
+Create Date: 2026-04-03
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+revision: str = "002_irm_core"
+down_revision: Union[str, None] = "001_initial"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS incidents (
+            id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+            title text NOT NULL,
+            status text NOT NULL DEFAULT 'open',
+            severity text NOT NULL DEFAULT 'warning',
+            source text NOT NULL DEFAULT 'grafana',
+            ingress_event_id uuid REFERENCES ingress_events (id) ON DELETE SET NULL,
+            created_at timestamptz NOT NULL DEFAULT now(),
+            updated_at timestamptz NOT NULL DEFAULT now()
+        );
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS incidents_created_at_idx
+        ON incidents (created_at DESC);
+        """
+    )
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS tasks (
+            id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+            incident_id uuid REFERENCES incidents (id) ON DELETE CASCADE,
+            title text NOT NULL,
+            status text NOT NULL DEFAULT 'open',
+            created_at timestamptz NOT NULL DEFAULT now()
+        );
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS tasks_incident_id_idx ON tasks (incident_id);
+        """
+    )
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS escalation_policies (
+            id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+            name text NOT NULL,
+            enabled boolean NOT NULL DEFAULT true,
+            steps jsonb NOT NULL DEFAULT '[]'::jsonb,
+            created_at timestamptz NOT NULL DEFAULT now()
+        );
+        """
+    )
+
+
+def downgrade() -> None:
+    op.execute("DROP TABLE IF EXISTS tasks;")
+    op.execute("DROP TABLE IF EXISTS incidents;")
+    op.execute("DROP TABLE IF EXISTS escalation_policies;")

alembic/versions/003_ingress_org_service.py

@@ -0,0 +1,48 @@
+"""ingress org/service + incident dimensions for multi-grafana
+
+Revision ID: 003_ingress_org
+Revises: 002_irm_core
+Create Date: 2026-04-03
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+revision: str = "003_ingress_org"
+down_revision: Union[str, None] = "002_irm_core"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute(
+        """
+        ALTER TABLE ingress_events
+        ADD COLUMN IF NOT EXISTS org_slug text,
+        ADD COLUMN IF NOT EXISTS service_name text;
+        """
+    )
+    op.execute(
+        """
+        ALTER TABLE incidents
+        ADD COLUMN IF NOT EXISTS grafana_org_slug text,
+        ADD COLUMN IF NOT EXISTS service_name text;
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS incidents_org_service_idx
+        ON incidents (grafana_org_slug, service_name)
+        WHERE grafana_org_slug IS NOT NULL OR service_name IS NOT NULL;
+        """
+    )
+
+
+def downgrade() -> None:
+    op.execute("DROP INDEX IF EXISTS incidents_org_service_idx;")
+    op.execute("ALTER TABLE incidents DROP COLUMN IF EXISTS service_name;")
+    op.execute("ALTER TABLE incidents DROP COLUMN IF EXISTS grafana_org_slug;")
+    op.execute("ALTER TABLE ingress_events DROP COLUMN IF EXISTS service_name;")
+    op.execute("ALTER TABLE ingress_events DROP COLUMN IF EXISTS org_slug;")

alembic/versions/004_grafana_catalog.py

@@ -0,0 +1,88 @@
+"""Grafana topology cache: org, folders, alert rules per instance slug
+
+Revision ID: 004_grafana_catalog
+Revises: 003_ingress_org
+Create Date: 2026-04-03
+
+"""
+
+from typing import Sequence, Union
+
+from alembic import op
+
+revision: str = "004_grafana_catalog"
+down_revision: Union[str, None] = "003_ingress_org"
+branch_labels: Union[str, Sequence[str], None] = None
+depends_on: Union[str, Sequence[str], None] = None
+
+
+def upgrade() -> None:
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS grafana_catalog_meta (
+            instance_slug text NOT NULL,
+            grafana_org_id int NOT NULL,
+            org_name text NOT NULL DEFAULT '',
+            synced_at timestamptz NOT NULL DEFAULT now(),
+            folder_count int NOT NULL DEFAULT 0,
+            rule_count int NOT NULL DEFAULT 0,
+            error_text text,
+            PRIMARY KEY (instance_slug, grafana_org_id)
+        );
+        """
+    )
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS grafana_catalog_folders (
+            id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+            instance_slug text NOT NULL,
+            grafana_org_id int NOT NULL,
+            folder_uid text NOT NULL,
+            title text NOT NULL DEFAULT '',
+            parent_uid text,
+            updated_at timestamptz NOT NULL DEFAULT now(),
+            UNIQUE (instance_slug, grafana_org_id, folder_uid)
+        );
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS grafana_cat_folders_inst_org_idx
+        ON grafana_catalog_folders (instance_slug, grafana_org_id);
+        """
+    )
+    op.execute(
+        """
+        CREATE TABLE IF NOT EXISTS grafana_catalog_rules (
+            id uuid PRIMARY KEY DEFAULT gen_random_uuid(),
+            instance_slug text NOT NULL,
+            grafana_org_id int NOT NULL,
+            namespace_uid text NOT NULL,
+            rule_group_name text NOT NULL,
+            rule_uid text NOT NULL,
+            title text NOT NULL DEFAULT '',
+            rule_group_interval text,
+            labels jsonb NOT NULL DEFAULT '{}'::jsonb,
+            updated_at timestamptz NOT NULL DEFAULT now(),
+            UNIQUE (instance_slug, grafana_org_id, rule_uid)
+        );
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS grafana_cat_rules_inst_org_idx
+        ON grafana_catalog_rules (instance_slug, grafana_org_id);
+        """
+    )
+    op.execute(
+        """
+        CREATE INDEX IF NOT EXISTS grafana_cat_rules_ns_idx
+        ON grafana_catalog_rules (instance_slug, grafana_org_id, namespace_uid);
+        """
+    )
+
+
+def downgrade() -> None:
+    op.execute("DROP TABLE IF EXISTS grafana_catalog_rules;")
+    op.execute("DROP TABLE IF EXISTS grafana_catalog_folders;")
+    op.execute("DROP TABLE IF EXISTS grafana_catalog_meta;")

deploy/entrypoint.sh

@@ -0,0 +1,7 @@
+#!/bin/sh
+set -e
+cd /app
+if [ -n "${DATABASE_URL:-}" ] && [ "${SKIP_ALEMBIC:-0}" != "1" ]; then
+  alembic upgrade head
+fi
+exec uvicorn onguard24.main:app --host 0.0.0.0 --port "${PORT:-8080}"

docker-compose.yml

@@ -0,0 +1,21 @@
+# Прод: на сервере рядом с репозиторием лежит .env (не в git).
+# Сборка: docker compose build && docker compose up -d
+services:
+  onguard24:
+    build: .
+    image: onguard24:latest
+    container_name: onguard24
+    restart: unless-stopped
+    env_file:
+      - .env
+    ports:
+      - "${ONGUARD_HTTP_PORT:-8080}:8080"
+    environment:
+      SKIP_ALEMBIC: ${SKIP_ALEMBIC:-0}
+      PORT: "8080"
+    healthcheck:
+      test: ["CMD", "python", "-c", "import urllib.request; urllib.request.urlopen('http://127.0.0.1:8080/health')"]
+      interval: 30s
+      timeout: 5s
+      retries: 3
+      start_period: 15s

docs/ARCHITECTURE.md

@@ -20,11 +20,15 @@ onGuard24/
 │   ├── integrations/
 │   │   ├── grafana_api.py   # Grafana HTTP API (Bearer SA)
 │   │   └── forgejo_api.py   # Forgejo/Gitea API (token + probe/fallback)
-│   └── modules/             # API + ui_router + registry + ui_support (фрагменты главной)
+│   └── modules/             # IRM: incidents, tasks, escalations, … + registry + ui_support
 ├── web/                     # Vite + React (опционально)
+├── Dockerfile
+├── docker-compose.yml
+├── deploy/entrypoint.sh
 ├── pyproject.toml
 ├── pytest.ini
 ├── tests/                   # pytest: health, status, ingress
+├── .gitea/workflows/        # CI + SSH deploy (Forgejo Actions)
 ├── CHANGELOG.md
 └── docs/
 ```

docs/CICD.md

@@ -0,0 +1,103 @@
+# CI/CD (Forgejo / Gitea) и деплой на `pvestandt9`
+
+Цель: **пуш тега `v*`** → автоматический деплой на сервер; **откат** — повторный запуск workflow с другим тегом.
+
+## Что в репозитории
+
+| Файл | Назначение |
+|------|------------|
+| `.gitea/workflows/ci.yml` | На `push` в `main` и PR: `pytest`. |
+| `.gitea/workflows/deploy.yml` | На `push` тега `v*`: SSH на сервер → `git checkout` → `docker compose build/up`. |
+| `Dockerfile` | Образ Python 3.12, `pip install .`, entrypoint: `alembic upgrade` + `uvicorn`. |
+| `docker-compose.yml` | Сервис `onguard24`, порт `8080`, `env_file: .env`. |
+| `deploy/entrypoint.sh` | Перед стартом: `alembic upgrade head` (отключить: `SKIP_ALEMBIC=1` в `.env`). |
+
+## Включить Actions в Forgejo
+
+1. Админка инстанса или репозитория: включить **Actions**.
+2. Зарегистрировать **runner** с меткой `ubuntu-latest` (или изменить `runs-on` в YAML на вашу метку, например `self-hosted`).
+3. Если образы `actions/checkout` недоступны, в настройках Actions задайте зеркало GitHub или используйте встроенные экшены Forgejo (см. документацию вашей версии).
+
+### Образ act_runner (Docker Hub)
+
+Используйте **`gitea/act_runner`** (например тег **`nightly`**), а не `docker.gitea.com/...:latest` — последний часто недоступен.
+
+```bash
+sudo docker pull gitea/act_runner:nightly
+sudo docker run -d --restart always --name act_runner \
+  -e GITEA_INSTANCE_URL="https://forgejo.pvenode.ru" \
+  -e GITEA_RUNNER_REGISTRATION_TOKEN="ТОКЕН_ИЗ_UI" \
+  -e GITEA_RUNNER_NAME="pvestandt9" \
+  -e GITEA_RUNNER_LABELS="ubuntu-latest:docker://catthehacker/ubuntu:act-22.04" \
+  -v /var/run/docker.sock:/var/run/docker.sock \
+  -v act_runner_data:/data \
+  gitea/act_runner:nightly
+```
+
+Стабильные теги: [hub.docker.com/r/gitea/act_runner/tags](https://hub.docker.com/r/gitea/act_runner/tags).
+
+## Секреты репозитория
+
+**Настройки репозитория → Actions → Secrets:**
+
+| Секрет | Пример | Описание |
+|--------|--------|----------|
+| `DEPLOY_HOST` | `pvestandt9` или IP | Хост SSH. |
+| `DEPLOY_USER` | `root` | Пользователь SSH. |
+| `DEPLOY_SSH_KEY` | содержимое `id_rsa` | Приватный ключ (весь PEM, многострочный). |
+| `DEPLOY_PATH` | `/opt/onGuard24` | Каталог **клона git** на сервере (там же `docker-compose.yml`). |
+
+Ключ лучше отдельный **deploy key** только на чтение репозитория или полный доступ, если runner делает только `git fetch`.
+
+## Однократная подготовка сервера (`root@pvestandt9`)
+
+```bash
+ssh root@pvestandt9
+apt-get update && apt-get install -y git docker.io docker-compose-v2
+# или Docker CE по инструкции вашей ОС
+
+mkdir -p /opt/onGuard24
+cd /opt
+git clone https://forgejo.pvenode.ru/admin/onGuard24.git onGuard24
+cd onGuard24
+```
+
+Создайте **`/opt/onGuard24/.env`** (как в `.env.example`): `DATABASE_URL`, `HTTP_ADDR` (в контейнере порт задаёт compose; для приложения можно `0.0.0.0:8080`), секреты Grafana и т.д.
+
+Настройте **`git remote`** и доступ (`~/.ssh` deploy key или `git credential`), чтобы **`git fetch` без пароля** работал от пользователя, под которым заходит CI (часто `root`).
+
+Проверка вручную:
+
+```bash
+cd /opt/onGuard24
+docker compose build && docker compose up -d
+curl -s http://127.0.0.1:8080/health
+```
+
+При необходимости пробросьте порт наружу (nginx, firewall).
+
+## Релиз (деплой новой версии)
+
+1. Закоммитить код, запушить тег: `git tag v1.6.0 && git push origin v1.6.0`.
+2. Workflow **Deploy** запустится сам, на сервере обновится код и пересоберётся контейнер.
+
+## Откат версии (простой процесс)
+
+1. В Forgejo: **Actions → Deploy → Run workflow**.
+2. В поле **ref** указать **старый тег**, например `v1.4.1`.
+3. Запустить. На сервере выполнится `checkout` и `reset` на этот ref, затем `docker compose build && up -d`.
+
+Убедитесь, что старый тег есть в `origin` (`git push --tags` не удалялся).
+
+## Миграции БД
+
+По умолчанию при каждом старте контейнера выполняется **`alembic upgrade head`**. Если нужно отключить (и гонять миграции вручную), в `.env` на сервере: `SKIP_ALEMBIC=1`.
+
+## Устранение неполадок
+
+- **Runner не берёт job** — проверьте `runs-on` и метки runner.
+- **SSH fails** — `ssh -i key root@DEPLOY_HOST` с машины runner; `known_hosts` при необходимости добавьте в экшен (расширение `appleboy/ssh-action` / `ssh-keyscan`).
+- **`git checkout` fails** — выполните на сервере `git fetch --tags` вручную, проверьте remote URL и ключ.
+- **База недоступна из контейнера** — в `DATABASE_URL` укажите хост, доступный **из Docker** (не `127.0.0.1` хоста, если БД на хосте — используйте IP хоста или `host.docker.internal` где поддерживается).
+
+См. также [VERSIONING.md](VERSIONING.md) и [IRM.md](IRM.md).

docs/GRAFANA_TOPOLOGY.md

@@ -0,0 +1,54 @@
+# Каталог Grafana в onGuard24
+
+Иерархия **инстанс → организация → папки → правила алертинга** подтягивается **по HTTP API** Grafana и сохраняется в PostgreSQL. Вебхук при этом не заменяет синхронизацию: вебхук даёт события, каталог — актуальную структуру правил.
+
+## Что нужно
+
+1. Миграция **`004_grafana_catalog`** (выполняется при старте контейнера через `alembic upgrade head`).
+2. В `.env` у источника Grafana заданы **`api_url`** и **`api_token`** (service account с правами читать папки и alert rules — обычно роль Viewer/Editor и доступ к Alerting).
+3. Источник попадает в **`iter_grafana_sources`**: `GRAFANA_SOURCES_JSON` или пара `GRAFANA_URL` + `GRAFANA_SERVICE_ACCOUNT_TOKEN` (slug `default`).
+
+## API
+
+| Метод | Путь | Назначение |
+|--------|------|------------|
+| POST | `/api/v1/modules/grafana-catalog/sync` | Синхронизация. Тело: `{}` — все источники с токеном; `{"instance_slug":"default"}` — один slug. |
+| GET | `/api/v1/modules/grafana-catalog/meta` | Последние метаданные синхронизации по всем slug. |
+| GET | `/api/v1/modules/grafana-catalog/tree?instance_slug=default` | Дерево: папки и правила, сгруппированные по `namespace_uid` (обычно UID папки правил). |
+
+## Проверка на https://onguard24.pvenode.ru/
+
+1. Убедитесь, что задеплоена версия с модулем **Каталог Grafana** (пункт в боковом меню).
+2. Выполните синхронизацию (с сервера или с машины с доступом к API):
+
+```bash
+curl -sS -X POST "https://onguard24.pvenode.ru/api/v1/modules/grafana-catalog/sync" \
+  -H "Content-Type: application/json" \
+  -d '{}'
+```
+
+Ответ: `results[]` с полями `ok`, `folders`, `rules`, при ошибке — `error`.
+
+3. Посмотреть дерево:
+
+```bash
+curl -sS "https://onguard24.pvenode.ru/api/v1/modules/grafana-catalog/tree?instance_slug=default" | python3 -m json.tool
+```
+
+(Если используете только `GRAFANA_URL`, slug источника — **`default`**.)
+
+4. В UI: **Каталог Grafana** — таблица последних синхронизаций.
+
+## Связь с вебхуком
+
+- В инцидентах по-прежнему используется **`externalURL` / лейблы** из тела вебхука.
+- Каталог позволяет в UI/API сопоставлять **`rule_uid`** и папку с сохранённым снимком (после доработок привязки).
+
+## Ограничения
+
+- Один токен = **одна Grafana-организация** (контекст `GET /api/org`). Обход нескольких org на одном инстансе — отдельные токены или админ-API (не в этой версии).
+- Эндпоинты Ruler: пробуются `/api/ruler/grafana/api/v1/rules` и `/api/ruler/Grafana/api/v1/rules` (зависит от версии Grafana).
+
+## Автоматический cron
+
+Периодический `POST .../sync` пока не встроен: можно повесить **системный cron**, **Forgejo scheduled workflow** или внешний оркестратор.

docs/IRM.md

@@ -0,0 +1,39 @@
+# IRM: функционал, назначение и реализация в onGuard24
+
+Краткий ориентир для разработки (аналог облачного IRM: инциденты, задачи, эскалации, дежурства).
+
+## Матрица: что это, зачем, как у нас, что в Grafana
+
+| Область | Назначение | onGuard24 | Grafana / внешнее |
+|---------|------------|-----------|-------------------|
+| **Инциденты** | Учёт сбоев, статусы (open → resolved), связь с алертом | Модуль `incidents`: таблица `incidents`, API, UI, авто-создание из `alert.received` | Contact point **Webhook** → `POST /api/v1/ingress/grafana`; правила алертинга в Grafana |
+| **Задачи** | Подзадачи по инциденту (разбор, фикс) | Модуль `tasks`: таблица `tasks`, привязка к `incident_id` | Опционально: ссылки из алерта; основная работа в onGuard24 |
+| **Цепочки эскалаций** | Кого звать и в каком порядке при таймаутах | Модуль `escalations`: таблица `escalation_policies` (JSON `steps`), API/UI заготовка | Маршрутизация уведомлений может дублироваться в Grafana contact points; целевая логика — в onGuard24 |
+| **Календарь дежурств** | Кто в смене, расписание | Модуль `schedules` (развитие) | Календари/команды — данные в onGuard24; уведомления — через интеграции |
+| **Контакты** | Люди, каналы | Модуль `contacts` | Получатели в **Contact points** (email, Slack, webhook) |
+| **Светофор / статус сервисов** | Агрегат здоровья | Модуль `statusboard` | Источник метрик — Prometheus/Loki; правила — Grafana |
+| **Группы алертов** | Группировка шумных алертов | *План:* отдельная сущность / правила | **Alertmanager** / группировка в Grafana Alerting |
+| **Интеграции** | Внешние системы | `integrations/`, статус в `/api/v1/status` | API-ключи Grafana, Vault, Forgejo в `.env` |
+| **Пользователи / права** | RBAC | *Пока нет* | SSO Grafana, сеть за reverse proxy |
+| **SLO** | Цели по доступности | *Вне скоупа v1* | Grafana SLO / Mimir |
+
+## Поток данных (алерт → инцидент)
+
+1. Grafana срабатывает правило → шлёт JSON на **webhook** onGuard24.
+2. Сервис пишет строку в `ingress_events`, публикует **`alert.received`**.
+3. Модуль **incidents** подписан на событие и создаёт запись в **`incidents`** с ссылкой на `ingress_event_id`.
+
+## Что настроить в Grafana (обязательно для приёма алертов)
+
+### Один URL для всех инстансов и организаций
+
+1. **Contact point → Webhook**, URL: **`https://<ваш-хост>/api/v1/ingress/grafana`** (POST). **Не нужно** заводить slug в `.env`: источник в БД определяется из JSON Grafana — **`externalURL`** (хост Grafana), при наличии **`orgId`** / **`org_id`**, иначе лейблы первого алерта (`__org_id__`, `grafana_org`, `tenant`, `cluster`, `namespace`).
+2. В **`grafana.ini`** / настройках сервера корректно задайте **`root_url` / `external URL`**, чтобы в вебхук попадал нужный хост (за NPM — публичный URL).
+3. Опционально **`X-OnGuard-Secret`** (если задан **`GRAFANA_WEBHOOK_SECRET`**) и **`X-OnGuard-Service`** (имя сервиса в инциденте).
+4. **Notification policies** — привязать правила к contact point.
+
+### Дополнительно: путь `/ingress/grafana/<slug>`
+
+Явный ярлык в URL **не требует** записи в `GRAFANA_SOURCES_JSON`. **`GRAFANA_SOURCES_JSON`** нужен в основном для **`/api/v1/status`** (проверка API каждого инстанса) и для **отдельного `webhook_secret` на slug**.
+
+Подробнее: [MODULES.md](MODULES.md), [DOMAIN.md](DOMAIN.md).

docs/MODULES.md

@@ -33,7 +33,7 @@
    - `router` — JSON API под `/api/v1/modules/<имя>/`.
    - Опционально **`ui_router`** — `APIRouter(include_in_schema=False)`, маршруты полных HTML-страниц (корень `/` → `/ui/modules/<slug>/`).
    - Опционально **`async def render_home_fragment(request) -> str`** — HTML-фрагмент (без `<html>`) для карточки на главной.
-   - **`register_events(_bus)`** — подписки на шину.
+   - **`register_events(bus, pool)`** — подписки на шину; при необходимости используйте **`pool`** для записи в БД из обработчика события.
 
 2. **Регистрация** в **`onguard24/modules/registry.py`** — объект **`ModuleMount`**:
    - `router`, `url_prefix`, `register_events`, **`slug`**, **`title`**, опционально **`ui_router`**, **`render_home_fragment`**.

onguard24/__init__.py

@@ -1,3 +1,3 @@
 """onGuard24 — модульный монолит (ядро + модули)."""
 
-__version__ = "1.4.0"
+__version__ = "1.7.0"

onguard24/config.py

@@ -19,6 +19,9 @@ class Settings(BaseSettings):
     http_addr: str = Field(default="0.0.0.0:8080", validation_alias="HTTP_ADDR")
     database_url: str = Field(default="", validation_alias="DATABASE_URL")
     grafana_webhook_secret: str = Field(default="", validation_alias="GRAFANA_WEBHOOK_SECRET")
+    # JSON-массив: [{"slug":"adibrov","api_url":"https://...","api_token":"glsa_...","webhook_secret":"..."}]
+    # Пусто + задан GRAFANA_URL → один источник slug "default"
+    grafana_sources_json: str = Field(default="", validation_alias="GRAFANA_SOURCES_JSON")
     # HTTP API (service account): Grafana → Administration → Service accounts → токен
     grafana_url: str = Field(default="", validation_alias="GRAFANA_URL")
     grafana_service_account_token: str = Field(

onguard24/deps.py

@@ -0,0 +1,9 @@
+"""Общие зависимости FastAPI."""
+
+from __future__ import annotations
+
+from fastapi import Request
+
+
+def get_pool(request: Request):
+    return getattr(request.app.state, "pool", None)

onguard24/domain/events.py

@@ -26,6 +26,8 @@ class AlertReceived(DomainEvent):
     name: str = "alert.received"
     alert: Alert | None = None
     raw_payload_ref: UUID | None = None
+    grafana_org_slug: str | None = None
+    service_name: str | None = None
 
 
 Handler = Callable[[DomainEvent], Awaitable[None]]
@@ -59,6 +61,18 @@ class InMemoryEventBus:
         for h in self._subs.get(event.name, []):
             await h(event)
 
-    async def publish_alert_received(self, alert: Alert, raw_payload_ref: UUID | None = None) -> None:
+    async def publish_alert_received(
-        ev = AlertReceived(alert=alert, raw_payload_ref=raw_payload_ref)
+        self,
+        alert: Alert,
+        raw_payload_ref: UUID | None = None,
+        *,
+        grafana_org_slug: str | None = None,
+        service_name: str | None = None,
+    ) -> None:
+        ev = AlertReceived(
+            alert=alert,
+            raw_payload_ref=raw_payload_ref,
+            grafana_org_slug=grafana_org_slug,
+            service_name=service_name,
+        )
         await self.publish(ev)

onguard24/grafana_sources.py

@@ -0,0 +1,84 @@
+"""Несколько инстансов Grafana: URL API, токен, секрет вебхука по slug (организация / стек)."""
+
+from __future__ import annotations
+
+import json
+import re
+from pydantic import BaseModel, Field, field_validator
+
+_SLUG_RE = re.compile(r"^[a-z0-9][a-z0-9_-]{0,62}$")
+
+
+class GrafanaSourceEntry(BaseModel):
+    """Один инстанс Grafana для API и/или приёма вебхуков."""
+
+    slug: str = Field(..., description="Идентификатор в URL: /ingress/grafana/{slug}")
+    api_url: str = Field(..., description="Базовый URL без завершающего слэша")
+    api_token: str = Field(default="", description="Service account token для HTTP API")
+    webhook_secret: str = Field(
+        default="",
+        description="Если задан — только этот секрет для вебхука этого slug; иначе см. GRAFANA_WEBHOOK_SECRET",
+    )
+
+    @field_validator("slug")
+    @classmethod
+    def slug_ok(cls, v: str) -> str:
+        s = v.strip().lower()
+        if not _SLUG_RE.match(s):
+            raise ValueError(
+                "slug: только a-z, цифры, - и _, длина 1–63, с буквы/цифры"
+            )
+        return s
+
+    @field_validator("api_url")
+    @classmethod
+    def strip_url(cls, v: str) -> str:
+        return v.rstrip("/")
+
+
+def parse_grafana_sources_json(raw: str) -> list[GrafanaSourceEntry]:
+    if not raw.strip():
+        return []
+    data = json.loads(raw)
+    if not isinstance(data, list):
+        raise ValueError("GRAFANA_SOURCES_JSON должен быть JSON-массивом объектов")
+    return [GrafanaSourceEntry.model_validate(x) for x in data]
+
+
+def iter_grafana_sources(settings: "Settings") -> list[GrafanaSourceEntry]:
+    """Список источников: из GRAFANA_SOURCES_JSON или один синтетический default из GRAFANA_URL."""
+    try:
+        parsed = parse_grafana_sources_json(settings.grafana_sources_json)
+    except (json.JSONDecodeError, ValueError):
+        parsed = []
+    if parsed:
+        return parsed
+    gu = settings.grafana_url.strip()
+    if not gu:
+        return []
+    return [
+        GrafanaSourceEntry(
+            slug="default",
+            api_url=gu.rstrip("/"),
+            api_token=settings.grafana_service_account_token.strip(),
+            webhook_secret="",
+        )
+    ]
+
+
+def sources_by_slug(settings: "Settings") -> dict[str, GrafanaSourceEntry]:
+    return {s.slug: s for s in iter_grafana_sources(settings)}
+
+
+def effective_webhook_secret(settings: "Settings", source: GrafanaSourceEntry | None) -> str:
+    """Пустая строка = проверка вебхука отключена (только для dev)."""
+    if source and source.webhook_secret.strip():
+        return source.webhook_secret.strip()
+    return settings.grafana_webhook_secret.strip()
+
+
+def webhook_authorized(settings: "Settings", source: GrafanaSourceEntry | None, header: str | None) -> bool:
+    need = effective_webhook_secret(settings, source)
+    if not need:
+        return True
+    return (header or "") == need

onguard24/ingress/grafana.py

@@ -1,29 +1,92 @@
 import json
 import logging
+import re
 from datetime import datetime, timezone
+from urllib.parse import urlparse
 
 from fastapi import APIRouter, Depends, Header, HTTPException, Request
 from starlette.responses import Response
 
 from onguard24.domain.entities import Alert, Severity
+from onguard24.grafana_sources import sources_by_slug, webhook_authorized
 
 logger = logging.getLogger(__name__)
 router = APIRouter(tags=["ingress"])
 
+_PATH_SLUG_RE = re.compile(r"^[a-z0-9][a-z0-9_-]{0,62}$")
+
+
+def sanitize_source_key(raw: str) -> str:
+    s = re.sub(r"[^a-zA-Z0-9._-]+", "-", raw.strip()).strip("-").lower()
+    return s[:200] if s else ""
+
+
+def extract_grafana_source_key(body: dict) -> str | None:
+    """
+    Идентификатор инстанса/организации из тела вебхука Grafana (без ручной прописки в .env).
+    См. поля payload: externalURL, orgId, лейблы в alerts[].
+    """
+    chunks: list[str] = []
+    ext = body.get("externalURL") or body.get("external_url")
+    if isinstance(ext, str) and ext.strip():
+        try:
+            host = urlparse(ext.strip()).hostname
+            if host:
+                chunks.append(host)
+        except Exception:
+            pass
+    for key in ("orgId", "org_id"):
+        v = body.get(key)
+        if v is not None and str(v).strip():
+            chunks.append(f"o{str(v).strip()[:24]}")
+            break
+    alerts = body.get("alerts")
+    if isinstance(alerts, list) and alerts:
+        a0 = alerts[0] if isinstance(alerts[0], dict) else {}
+        labels = a0.get("labels") if isinstance(a0.get("labels"), dict) else {}
+        for lk in ("__org_id__", "grafana_org", "tenant", "cluster", "namespace"):
+            v = labels.get(lk)
+            if v is not None and str(v).strip():
+                chunks.append(str(v).strip()[:64])
+                break
+    if not chunks:
+        return None
+    return sanitize_source_key("-".join(chunks)) or None
+
 
 async def get_pool(request: Request):
     return getattr(request.app.state, "pool", None)
 
 
-@router.post("/ingress/grafana", status_code=202)
+def service_hint_from_grafana_body(body: dict, header_service: str | None) -> str | None:
-async def grafana_webhook(
+    """Имя сервиса: заголовок X-OnGuard-Service или лейблы из Unified Alerting."""
+    if header_service and header_service.strip():
+        return header_service.strip()[:200]
+    alerts = body.get("alerts")
+    if isinstance(alerts, list) and alerts:
+        labels = alerts[0].get("labels") if isinstance(alerts[0], dict) else None
+        if isinstance(labels, dict):
+            for key in ("service", "job", "namespace", "cluster", "instance"):
+                v = labels.get(key)
+                if v is not None and str(v).strip():
+                    return str(v).strip()[:200]
+    common = body.get("commonLabels")
+    if isinstance(common, dict):
+        for key in ("service", "job", "namespace"):
+            v = common.get(key)
+            if v is not None and str(v).strip():
+                return str(v).strip()[:200]
+    return None
+
+
+async def _grafana_webhook_impl(
     request: Request,
-    pool=Depends(get_pool),
+    pool,
-    x_onguard_secret: str | None = Header(default=None, alias="X-OnGuard-Secret"),
+    x_onguard_secret: str | None,
-):
+    x_onguard_service: str | None,
+    path_slug: str | None,
+) -> Response:
     settings = request.app.state.settings
-    if settings.grafana_webhook_secret and x_onguard_secret != settings.grafana_webhook_secret:
-        raise HTTPException(status_code=401, detail="unauthorized")
 
     raw = await request.body()
     if len(raw) > 1_000_000:
@@ -32,6 +95,25 @@ async def grafana_webhook(
         body = json.loads(raw.decode() or "{}")
     except json.JSONDecodeError:
         body = {}
+    if not isinstance(body, dict):
+        body = {}
+
+    derived = extract_grafana_source_key(body)
+    path_key: str | None = None
+    if path_slug is not None:
+        path_key = path_slug.strip().lower()
+        if not _PATH_SLUG_RE.match(path_key):
+            raise HTTPException(status_code=400, detail="invalid path slug")
+        stored_org_slug = path_key
+    else:
+        stored_org_slug = derived
+
+    by = sources_by_slug(settings)
+    source = by.get(path_key) if path_key else None
+    if not webhook_authorized(settings, source, x_onguard_secret):
+        raise HTTPException(status_code=401, detail="unauthorized")
+
+    service_name = service_hint_from_grafana_body(body, x_onguard_service)
 
     if pool is None:
         logger.warning("ingress: database not configured, event not persisted")
@@ -39,9 +121,15 @@ async def grafana_webhook(
 
     async with pool.acquire() as conn:
         row = await conn.fetchrow(
-            "INSERT INTO ingress_events (source, body) VALUES ($1, $2::jsonb) RETURNING id",
+            """
+            INSERT INTO ingress_events (source, body, org_slug, service_name)
+            VALUES ($1, $2::jsonb, $3, $4)
+            RETURNING id
+            """,
             "grafana",
             json.dumps(body),
+            stored_org_slug,
+            service_name,
         )
     raw_id = row["id"] if row else None
     bus = getattr(request.app.state, "event_bus", None)
@@ -54,5 +142,43 @@ async def grafana_webhook(
             payload=body,
             received_at=datetime.now(timezone.utc),
         )
-        await bus.publish_alert_received(alert, raw_payload_ref=raw_id)
+        await bus.publish_alert_received(
+            alert,
+            raw_payload_ref=raw_id,
+            grafana_org_slug=stored_org_slug,
+            service_name=service_name,
+        )
     return Response(status_code=202)
+
+
+@router.post("/ingress/grafana", status_code=202)
+async def grafana_webhook_legacy(
+    request: Request,
+    pool=Depends(get_pool),
+    x_onguard_secret: str | None = Header(default=None, alias="X-OnGuard-Secret"),
+    x_onguard_service: str | None = Header(default=None, alias="X-OnGuard-Service"),
+):
+    """
+    Универсальный URL для любого инстанса Grafana: org_slug в БД берётся из тела
+    (externalURL, orgId, лейблы), без преднастройки в .env.
+    """
+    return await _grafana_webhook_impl(
+        request, pool, x_onguard_secret, x_onguard_service, path_slug=None
+    )
+
+
+@router.post("/ingress/grafana/{org_slug}", status_code=202)
+async def grafana_webhook_org(
+    org_slug: str,
+    request: Request,
+    pool=Depends(get_pool),
+    x_onguard_secret: str | None = Header(default=None, alias="X-OnGuard-Secret"),
+    x_onguard_service: str | None = Header(default=None, alias="X-OnGuard-Service"),
+):
+    """
+    Опционально: явный ярлык в URL (перекрывает авто-извлечение из JSON).
+    Секрет для пути: webhook_secret из GRAFANA_SOURCES_JSON для этого slug, иначе общий.
+    """
+    return await _grafana_webhook_impl(
+        request, pool, x_onguard_secret, x_onguard_service, path_slug=org_slug
+    )

onguard24/integrations/grafana_topology.py

@@ -0,0 +1,169 @@
+"""Чтение иерархии Grafana: org, папки, managed alert rules (Ruler API)."""
+
+from __future__ import annotations
+
+import json
+import logging
+from dataclasses import dataclass
+from typing import Any
+
+import httpx
+
+log = logging.getLogger(__name__)
+
+_HEADERS_JSON = {"Accept": "application/json"}
+
+
+def _auth_headers(token: str) -> dict[str, str]:
+    return {**_HEADERS_JSON, "Authorization": f"Bearer {token.strip()}"}
+
+
+@dataclass
+class ParsedRuleRow:
+    namespace_uid: str
+    rule_group_name: str
+    rule_group_interval: str | None
+    rule_uid: str
+    title: str
+    labels: dict[str, Any]
+
+
+async def fetch_org(base_url: str, token: str) -> tuple[dict[str, Any] | None, str | None]:
+    base = base_url.rstrip("/")
+    try:
+        async with httpx.AsyncClient(timeout=30.0, verify=True, follow_redirects=True) as client:
+            r = await client.get(f"{base}/api/org", headers=_auth_headers(token))
+    except Exception as e:
+        return None, str(e)
+    if r.status_code != 200:
+        return None, f"http {r.status_code}: {(r.text or '')[:300]}"
+    try:
+        return r.json(), None
+    except Exception:
+        return None, "invalid json from /api/org"
+
+
+async def fetch_all_folders(base_url: str, token: str) -> tuple[list[dict[str, Any]], str | None]:
+    """Обход дерева папок через /api/folders?parentUid=…"""
+    base = base_url.rstrip("/")
+    out: list[dict[str, Any]] = []
+    seen: set[str] = set()
+    queue: list[str | None] = [None]
+
+    try:
+        async with httpx.AsyncClient(timeout=60.0, verify=True, follow_redirects=True) as client:
+            while queue:
+                parent_uid = queue.pop(0)
+                params: dict[str, str] = {}
+                if parent_uid is not None:
+                    params["parentUid"] = parent_uid
+                r = await client.get(f"{base}/api/folders", headers=_auth_headers(token), params=params)
+                if r.status_code != 200:
+                    return out, f"folders http {r.status_code}: {(r.text or '')[:200]}"
+                try:
+                    chunk = r.json()
+                except Exception:
+                    return out, "invalid json from /api/folders"
+                if not isinstance(chunk, list):
+                    return out, "folders response is not a list"
+                for f in chunk:
+                    if not isinstance(f, dict):
+                        continue
+                    uid = f.get("uid")
+                    if not uid or uid in seen:
+                        continue
+                    seen.add(str(uid))
+                    out.append(f)
+                    queue.append(str(uid))
+    except Exception as e:
+        return out, str(e)
+    return out, None
+
+
+async def fetch_ruler_rules_raw(base_url: str, token: str) -> tuple[dict[str, Any] | None, str | None]:
+    """GET Ruler API для Grafana-managed правил (namespace → группы → rules)."""
+    base = base_url.rstrip("/")
+    paths = (
+        "/api/ruler/grafana/api/v1/rules",
+        "/api/ruler/Grafana/api/v1/rules",
+    )
+    last_err: str | None = None
+    try:
+        async with httpx.AsyncClient(timeout=90.0, verify=True, follow_redirects=True) as client:
+            for path in paths:
+                r = await client.get(f"{base}{path}", headers=_auth_headers(token))
+                if r.status_code == 200:
+                    try:
+                        data = r.json()
+                    except Exception:
+                        return None, "invalid json from ruler"
+                    if isinstance(data, dict):
+                        return data, None
+                    return None, "ruler response is not an object"
+                last_err = f"ruler {path} http {r.status_code}: {(r.text or '')[:200]}"
+    except Exception as e:
+        return None, str(e)
+    return None, last_err or "ruler: no path matched"
+
+
+def parse_ruler_rules(data: dict[str, Any]) -> list[ParsedRuleRow]:
+    rows: list[ParsedRuleRow] = []
+    for namespace_uid, groups in data.items():
+        if not isinstance(namespace_uid, str) or not namespace_uid.strip():
+            continue
+        ns = namespace_uid.strip()
+        if not isinstance(groups, list):
+            continue
+        for grp in groups:
+            if not isinstance(grp, dict):
+                continue
+            gname = str(grp.get("name") or "group")
+            interval = grp.get("interval")
+            interval_s = str(interval) if interval is not None else None
+            rules = grp.get("rules")
+            if not isinstance(rules, list):
+                continue
+            for rule in rules:
+                if not isinstance(rule, dict):
+                    continue
+                ga = rule.get("grafana_alert")
+                if not isinstance(ga, dict):
+                    continue
+                uid = ga.get("uid") or rule.get("uid")
+                if not uid:
+                    continue
+                title = str(ga.get("title") or gname or uid)
+                labels = rule.get("labels")
+                if not isinstance(labels, dict):
+                    labels = {}
+                lbl = {str(k): str(v) for k, v in labels.items() if v is not None}
+                rows.append(
+                    ParsedRuleRow(
+                        namespace_uid=ns,
+                        rule_group_name=gname,
+                        rule_group_interval=interval_s,
+                        rule_uid=str(uid),
+                        title=title[:500],
+                        labels=lbl,
+                    )
+                )
+    return rows
+
+
+def merge_folder_rows(
+    api_folders: list[dict[str, Any]],
+    rule_namespaces: set[str],
+) -> list[tuple[str, str, str | None]]:
+    """(uid, title, parent_uid); добавляем namespace из ruler без записи в /api/folders."""
+    by_uid: dict[str, tuple[str, str | None]] = {}
+    for f in api_folders:
+        uid = f.get("uid")
+        if not uid:
+            continue
+        p = f.get("parentUid")
+        parent = str(p) if p else None
+        by_uid[str(uid)] = (str(f.get("title") or uid), parent)
+    for ns in rule_namespaces:
+        if ns not in by_uid:
+            by_uid[ns] = (ns, None)
+    return [(uid, t[0], t[1]) for uid, t in sorted(by_uid.items(), key=lambda x: x[0])]

onguard24/main.py

@@ -34,7 +34,7 @@ async def lifespan(app: FastAPI):
     settings = get_settings()
     pool = await create_pool(settings)
     bus = InMemoryEventBus()
-    register_module_events(bus)
+    register_module_events(bus, pool)
     app.state.pool = pool
     app.state.settings = settings
     app.state.event_bus = bus

onguard24/modules/contacts.py

@@ -9,7 +9,7 @@ router = APIRouter(tags=["module-contacts"])
 ui_router = APIRouter(tags=["web-contacts"], include_in_schema=False)
 
 
-def register_events(_bus: EventBus) -> None:
+def register_events(_bus: EventBus, _pool=None) -> None:
     pass
 
 

onguard24/modules/escalations.py

@@ -0,0 +1,220 @@
+"""IRM: цепочки эскалаций (политики в JSON, дальше — исполнение по шагам)."""
+
+from __future__ import annotations
+
+import html
+import json
+from uuid import UUID
+
+import asyncpg
+from fastapi import APIRouter, Depends, HTTPException, Request
+from fastapi.responses import HTMLResponse
+from pydantic import BaseModel, Field
+
+from onguard24.deps import get_pool
+from onguard24.domain.events import EventBus
+from onguard24.modules.ui_support import wrap_module_html_page
+
+router = APIRouter(tags=["module-escalations"])
+ui_router = APIRouter(tags=["web-escalations"], include_in_schema=False)
+
+
+class PolicyCreate(BaseModel):
+    name: str = Field(..., min_length=1, max_length=200)
+    enabled: bool = True
+    steps: list[dict] = Field(default_factory=list)
+
+
+class PolicyPatch(BaseModel):
+    name: str | None = Field(default=None, min_length=1, max_length=200)
+    enabled: bool | None = None
+    steps: list[dict] | None = None
+
+
+def register_events(_bus: EventBus, _pool: asyncpg.Pool | None = None) -> None:
+    pass
+
+
+async def render_home_fragment(request: Request) -> str:
+    pool = get_pool(request)
+    if pool is None:
+        return '<p class="module-note">Нужна БД для политик эскалации.</p>'
+    try:
+        async with pool.acquire() as conn:
+            n = await conn.fetchval("SELECT count(*)::int FROM escalation_policies WHERE enabled = true")
+    except Exception:
+        return '<p class="module-note">Таблица политик недоступна (миграции?).</p>'
+    return f'<div class="module-fragment"><p>Активных политик: <strong>{int(n)}</strong></p></div>'
+
+
+@router.get("/")
+async def list_policies_api(pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        return {"items": [], "database": "disabled"}
+    async with pool.acquire() as conn:
+        rows = await conn.fetch(
+            """
+            SELECT id, name, enabled, steps, created_at
+            FROM escalation_policies
+            ORDER BY name
+            """
+        )
+    items = []
+    for r in rows:
+        steps = r["steps"]
+        if isinstance(steps, str):
+            steps = json.loads(steps)
+        items.append(
+            {
+                "id": str(r["id"]),
+                "name": r["name"],
+                "enabled": r["enabled"],
+                "steps": steps if isinstance(steps, list) else [],
+                "created_at": r["created_at"].isoformat() if r["created_at"] else None,
+            }
+        )
+    return {"items": items}
+
+
+@router.post("/", status_code=201)
+async def create_policy_api(body: PolicyCreate, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            INSERT INTO escalation_policies (name, enabled, steps)
+            VALUES ($1, $2, $3::jsonb)
+            RETURNING id, name, enabled, steps, created_at
+            """,
+            body.name.strip(),
+            body.enabled,
+            json.dumps(body.steps),
+        )
+    steps = row["steps"]
+    return {
+        "id": str(row["id"]),
+        "name": row["name"],
+        "enabled": row["enabled"],
+        "steps": list(steps) if steps else [],
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+    }
+
+
+def _policy_dict(row) -> dict:
+    steps = row["steps"]
+    if isinstance(steps, str):
+        steps = json.loads(steps)
+    return {
+        "id": str(row["id"]),
+        "name": row["name"],
+        "enabled": row["enabled"],
+        "steps": steps if isinstance(steps, list) else [],
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+    }
+
+
+@router.get("/{policy_id}")
+async def get_policy_api(policy_id: UUID, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            SELECT id, name, enabled, steps, created_at
+            FROM escalation_policies WHERE id = $1::uuid
+            """,
+            policy_id,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return _policy_dict(row)
+
+
+@router.patch("/{policy_id}")
+async def patch_policy_api(
+    policy_id: UUID,
+    body: PolicyPatch,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    if body.name is None and body.enabled is None and body.steps is None:
+        raise HTTPException(status_code=400, detail="no fields to update")
+    steps_json = json.dumps(body.steps) if body.steps is not None else None
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            UPDATE escalation_policies SET
+                name = COALESCE($2, name),
+                enabled = COALESCE($3, enabled),
+                steps = COALESCE($4::jsonb, steps)
+            WHERE id = $1::uuid
+            RETURNING id, name, enabled, steps, created_at
+            """,
+            policy_id,
+            body.name.strip() if body.name is not None else None,
+            body.enabled,
+            steps_json,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return _policy_dict(row)
+
+
+@router.delete("/{policy_id}", status_code=204)
+async def delete_policy_api(policy_id: UUID, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            "DELETE FROM escalation_policies WHERE id = $1::uuid RETURNING id",
+            policy_id,
+        )
+    if row is None:
+        raise HTTPException(status_code=404, detail="not found")
+
+
+@ui_router.get("/", response_class=HTMLResponse)
+async def escalations_ui_home(request: Request):
+    pool = get_pool(request)
+    rows_html = ""
+    err = ""
+    if pool is None:
+        err = "<p>База данных не настроена.</p>"
+    else:
+        try:
+            async with pool.acquire() as conn:
+                rows = await conn.fetch(
+                    "SELECT id, name, enabled, steps FROM escalation_policies ORDER BY name"
+                )
+            for r in rows:
+                steps = r["steps"]
+                if hasattr(steps, "__iter__") and not isinstance(steps, (str, bytes)):
+                    steps_preview = html.escape(json.dumps(steps, ensure_ascii=False)[:120])
+                else:
+                    steps_preview = "—"
+                rows_html += (
+                    "<tr>"
+                    f"<td>{html.escape(str(r['id']))[:8]}…</td>"
+                    f"<td>{html.escape(r['name'])}</td>"
+                    f"<td>{'да' if r['enabled'] else 'нет'}</td>"
+                    f"<td><code>{steps_preview}</code></td>"
+                    "</tr>"
+                )
+        except Exception as e:
+            err = f"<p class=\"module-err\">{html.escape(str(e))}</p>"
+    inner = f"""<h1>Цепочки эскалаций</h1>
+<p>Заготовка: шаги хранятся в JSON; исполнение по таймерам — следующие версии.</p>
+{err}
+<table class="irm-table">
+<thead><tr><th>ID</th><th>Имя</th><th>Вкл.</th><th>Шаги (фрагмент)</th></tr></thead>
+<tbody>{rows_html or '<tr><td colspan="4">Нет политик — создайте через API POST</td></tr>'}</tbody>
+</table>"""
+    return HTMLResponse(
+        wrap_module_html_page(
+            document_title="Эскалации — onGuard24",
+            current_slug="escalations",
+            main_inner_html=inner,
+        )
+    )

onguard24/modules/grafana_catalog.py

@@ -0,0 +1,380 @@
+"""Кэш иерархии Grafana (инстанс → org → папки → правила) через API + БД."""
+
+from __future__ import annotations
+
+import html
+import json
+import logging
+from dataclasses import dataclass
+
+import asyncpg
+from fastapi import APIRouter, Depends, HTTPException, Request
+from fastapi.responses import HTMLResponse
+from pydantic import BaseModel, Field
+
+from onguard24.config import get_settings
+from onguard24.deps import get_pool
+from onguard24.domain.events import EventBus
+from onguard24.grafana_sources import iter_grafana_sources
+from onguard24.integrations import grafana_topology as gt
+from onguard24.modules.ui_support import wrap_module_html_page
+
+log = logging.getLogger(__name__)
+
+router = APIRouter(tags=["module-grafana-catalog"])
+ui_router = APIRouter(tags=["web-grafana-catalog"], include_in_schema=False)
+
+
+def register_events(_bus: EventBus, _pool: asyncpg.Pool | None = None) -> None:
+    pass
+
+
+class SyncBody(BaseModel):
+    instance_slug: str | None = Field(
+        default=None,
+        description="Slug из GRAFANA_SOURCES_JSON; пусто — все источники с api_token",
+    )
+
+
+@dataclass
+class PullOutcome:
+    org_id: int
+    org_name: str
+    folder_rows: list[tuple[str, str, str | None]]
+    rules: list[gt.ParsedRuleRow]
+    warnings: list[str]
+
+
+async def pull_topology(api_url: str, token: str) -> tuple[PullOutcome | None, str | None]:
+    org, oerr = await gt.fetch_org(api_url, token)
+    if oerr or not org:
+        return None, oerr or "no org"
+    oid = org.get("id")
+    if oid is None:
+        return None, "org response without id"
+    oname = str(org.get("name") or "")
+
+    warnings: list[str] = []
+    folders_raw, ferr = await gt.fetch_all_folders(api_url, token)
+    if ferr:
+        warnings.append(ferr)
+
+    ruler_raw, rerr = await gt.fetch_ruler_rules_raw(api_url, token)
+    if rerr:
+        warnings.append(rerr)
+        rules: list[gt.ParsedRuleRow] = []
+        namespaces: set[str] = set()
+    else:
+        rules = gt.parse_ruler_rules(ruler_raw or {})
+        namespaces = {r.namespace_uid for r in rules}
+
+    merged = gt.merge_folder_rows(folders_raw, namespaces)
+    return (
+        PullOutcome(
+            org_id=int(oid),
+            org_name=oname,
+            folder_rows=merged,
+            rules=rules,
+            warnings=warnings,
+        ),
+        None,
+    )
+
+
+async def persist_topology(
+    conn: asyncpg.Connection,
+    instance_slug: str,
+    outcome: PullOutcome,
+) -> None:
+    oid = outcome.org_id
+    await conn.execute(
+        """
+        DELETE FROM grafana_catalog_rules
+        WHERE instance_slug = $1 AND grafana_org_id = $2
+        """,
+        instance_slug,
+        oid,
+    )
+    await conn.execute(
+        """
+        DELETE FROM grafana_catalog_folders
+        WHERE instance_slug = $1 AND grafana_org_id = $2
+        """,
+        instance_slug,
+        oid,
+    )
+
+    for uid, title, parent in outcome.folder_rows:
+        await conn.execute(
+            """
+            INSERT INTO grafana_catalog_folders
+            (instance_slug, grafana_org_id, folder_uid, title, parent_uid)
+            VALUES ($1, $2, $3, $4, $5)
+            """,
+            instance_slug,
+            oid,
+            uid,
+            title,
+            parent,
+        )
+
+    for r in outcome.rules:
+        await conn.execute(
+            """
+            INSERT INTO grafana_catalog_rules (
+                instance_slug, grafana_org_id, namespace_uid, rule_group_name,
+                rule_uid, title, rule_group_interval, labels
+            )
+            VALUES ($1, $2, $3, $4, $5, $6, $7, $8::jsonb)
+            """,
+            instance_slug,
+            oid,
+            r.namespace_uid,
+            r.rule_group_name,
+            r.rule_uid,
+            r.title,
+            r.rule_group_interval,
+            json.dumps(r.labels),
+        )
+
+    fc = len(outcome.folder_rows)
+    rc = len(outcome.rules)
+    warn_txt = "; ".join(outcome.warnings) if outcome.warnings else None
+    if warn_txt and len(warn_txt) > 1900:
+        warn_txt = warn_txt[:1900] + "…"
+    await conn.execute(
+        """
+        INSERT INTO grafana_catalog_meta (
+            instance_slug, grafana_org_id, org_name, synced_at,
+            folder_count, rule_count, error_text
+        )
+        VALUES ($1, $2, $3, now(), $4, $5, $6)
+        ON CONFLICT (instance_slug, grafana_org_id) DO UPDATE SET
+            org_name = EXCLUDED.org_name,
+            synced_at = EXCLUDED.synced_at,
+            folder_count = EXCLUDED.folder_count,
+            rule_count = EXCLUDED.rule_count,
+            error_text = EXCLUDED.error_text
+        """,
+        instance_slug,
+        oid,
+        outcome.org_name,
+        fc,
+        rc,
+        warn_txt,
+    )
+
+
+@router.post("/sync", status_code=200)
+async def sync_catalog_api(
+    body: SyncBody,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    sources = iter_grafana_sources(get_settings())
+    if body.instance_slug:
+        sl = body.instance_slug.strip().lower()
+        sources = [s for s in sources if s.slug == sl]
+        if not sources:
+            raise HTTPException(status_code=404, detail="unknown instance slug")
+    to_run = [s for s in sources if s.api_token.strip()]
+    if not to_run:
+        raise HTTPException(
+            status_code=400,
+            detail="нет источников с api_token; задайте GRAFANA_SOURCES_JSON или GRAFANA_SERVICE_ACCOUNT_TOKEN",
+        )
+
+    results: list[dict] = []
+    for src in to_run:
+        outcome, err = await pull_topology(src.api_url, src.api_token)
+        if err or outcome is None:
+            results.append({"slug": src.slug, "ok": False, "error": err})
+            log.warning("grafana_catalog sync failed %s: %s", src.slug, err)
+            continue
+        async with pool.acquire() as conn:
+            async with conn.transaction():
+                await persist_topology(conn, src.slug, outcome)
+        results.append(
+            {
+                "slug": src.slug,
+                "ok": True,
+                "org_id": outcome.org_id,
+                "org_name": outcome.org_name,
+                "folders": len(outcome.folder_rows),
+                "rules": len(outcome.rules),
+                "warnings": outcome.warnings,
+            }
+        )
+    return {"results": results}
+
+
+@router.get("/meta")
+async def list_meta_api(pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        return {"items": [], "database": "disabled"}
+    async with pool.acquire() as conn:
+        rows = await conn.fetch(
+            """
+            SELECT instance_slug, grafana_org_id, org_name, synced_at,
+                   folder_count, rule_count, error_text
+            FROM grafana_catalog_meta
+            ORDER BY instance_slug, grafana_org_id
+            """
+        )
+    items = []
+    for r in rows:
+        items.append(
+            {
+                "instance_slug": r["instance_slug"],
+                "grafana_org_id": r["grafana_org_id"],
+                "org_name": r["org_name"],
+                "synced_at": r["synced_at"].isoformat() if r["synced_at"] else None,
+                "folder_count": r["folder_count"],
+                "rule_count": r["rule_count"],
+                "error_text": r["error_text"],
+            }
+        )
+    return {"items": items}
+
+
+@router.get("/tree")
+async def tree_api(
+    instance_slug: str,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    slug = instance_slug.strip().lower()
+    async with pool.acquire() as conn:
+        meta = await conn.fetchrow(
+            """
+            SELECT * FROM grafana_catalog_meta
+            WHERE instance_slug = $1 AND grafana_org_id > 0
+            ORDER BY synced_at DESC LIMIT 1
+            """,
+            slug,
+        )
+        if not meta:
+            raise HTTPException(status_code=404, detail="no catalog for this slug; run POST /sync first")
+        oid = meta["grafana_org_id"]
+        folders = await conn.fetch(
+            """
+            SELECT folder_uid, title, parent_uid
+            FROM grafana_catalog_folders
+            WHERE instance_slug = $1 AND grafana_org_id = $2
+            ORDER BY title
+            """,
+            slug,
+            oid,
+        )
+        rules = await conn.fetch(
+            """
+            SELECT namespace_uid, rule_group_name, rule_uid, title,
+                   rule_group_interval, labels
+            FROM grafana_catalog_rules
+            WHERE instance_slug = $1 AND grafana_org_id = $2
+            ORDER BY namespace_uid, rule_group_name, title
+            """,
+            slug,
+            oid,
+        )
+
+    by_ns: dict[str, list[dict]] = {}
+    for r in rules:
+        ns = r["namespace_uid"]
+        by_ns.setdefault(ns, []).append(
+            {
+                "rule_uid": r["rule_uid"],
+                "title": r["title"],
+                "rule_group": r["rule_group_name"],
+                "interval": r["rule_group_interval"],
+                "labels": r["labels"] if isinstance(r["labels"], dict) else {},
+            }
+        )
+
+    folder_nodes = []
+    for f in folders:
+        uid = f["folder_uid"]
+        folder_nodes.append(
+            {
+                "folder_uid": uid,
+                "title": f["title"],
+                "parent_uid": f["parent_uid"],
+                "rules": by_ns.get(uid, []),
+            }
+        )
+
+    return {
+        "instance_slug": slug,
+        "grafana_org_id": oid,
+        "org_name": meta["org_name"],
+        "synced_at": meta["synced_at"].isoformat() if meta["synced_at"] else None,
+        "folders": folder_nodes,
+        "orphan_rule_namespaces": sorted(set(by_ns.keys()) - {f["folder_uid"] for f in folders}),
+    }
+
+
+@ui_router.get("/", response_class=HTMLResponse)
+async def grafana_catalog_ui(request: Request):
+    pool = get_pool(request)
+    inner = ""
+    if pool is None:
+        inner = "<p>База не настроена.</p>"
+    else:
+        try:
+            async with pool.acquire() as conn:
+                rows = await conn.fetch(
+                    """
+                    SELECT instance_slug, org_name, synced_at, folder_count, rule_count, error_text
+                    FROM grafana_catalog_meta
+                    ORDER BY instance_slug
+                    """
+                )
+            if not rows:
+                inner = "<p>Каталог пуст. Вызовите <code>POST /api/v1/modules/grafana-catalog/sync</code>.</p>"
+            else:
+                inner = "<table class='irm-table'><thead><tr><th>Slug</th><th>Org</th><th>Синхр.</th><th>Папок</th><th>Правил</th><th>Ошибка</th></tr></thead><tbody>"
+                for r in rows:
+                    err = html.escape(str(r["error_text"] or "—"))[:120]
+                    st = r["synced_at"].isoformat() if r["synced_at"] else "—"
+                    inner += (
+                        f"<tr><td>{html.escape(r['instance_slug'])}</td>"
+                        f"<td>{html.escape(str(r['org_name']))}</td>"
+                        f"<td>{html.escape(st)}</td>"
+                        f"<td>{r['folder_count']}</td><td>{r['rule_count']}</td>"
+                        f"<td>{err}</td></tr>"
+                    )
+                inner += "</tbody></table>"
+        except Exception as e:
+            inner = f"<p class='module-err'>{html.escape(str(e))}</p>"
+    page = f"""<h1>Каталог Grafana</h1>
+<p>Иерархия: инстанс (slug) → организация → папки → правила. Синхронизация по HTTP API.</p>
+{inner}
+<p><small>API: <code>POST …/grafana-catalog/sync</code>, <code>GET …/grafana-catalog/tree?instance_slug=…</code></small></p>"""
+    return HTMLResponse(
+        wrap_module_html_page(
+            document_title="Каталог Grafana — onGuard24",
+            current_slug="grafana-catalog",
+            main_inner_html=page,
+        )
+    )
+
+
+async def render_home_fragment(request: Request) -> str:
+    pool = get_pool(request)
+    if pool is None:
+        return '<p class="module-note">Нужна БД для каталога Grafana.</p>'
+    try:
+        async with pool.acquire() as conn:
+            n = await conn.fetchval("SELECT count(*)::int FROM grafana_catalog_meta WHERE grafana_org_id >= 0")
+            last = await conn.fetchrow(
+                "SELECT max(synced_at) AS m FROM grafana_catalog_meta WHERE grafana_org_id >= 0"
+            )
+    except Exception:
+        return '<p class="module-note">Таблицы каталога недоступны (миграции?).</p>'
+    ts = last["m"].isoformat() if last and last["m"] else "никогда"
+    return (
+        f'<div class="module-fragment"><p>Источников с синхронизацией: <strong>{int(n)}</strong>. '
+        f"Последняя синхр.: <strong>{html.escape(ts)}</strong></p></div>"
+    )

onguard24/modules/incidents.py

@@ -0,0 +1,311 @@
+"""IRM: инциденты — учёт сбоев, связь с сырым ingress и событием alert.received."""
+
+from __future__ import annotations
+
+import html
+import logging
+from uuid import UUID
+
+import asyncpg
+from fastapi import APIRouter, Depends, HTTPException, Request
+from fastapi.responses import HTMLResponse
+from pydantic import BaseModel, Field
+
+from onguard24.deps import get_pool
+from onguard24.domain.events import AlertReceived, DomainEvent, EventBus
+from onguard24.modules.ui_support import wrap_module_html_page
+
+log = logging.getLogger(__name__)
+
+router = APIRouter(tags=["module-incidents"])
+ui_router = APIRouter(tags=["web-incidents"], include_in_schema=False)
+
+
+class IncidentCreate(BaseModel):
+    title: str = Field(..., min_length=1, max_length=500)
+    status: str = Field(default="open", max_length=64)
+    severity: str = Field(default="warning", max_length=32)
+
+
+class IncidentPatch(BaseModel):
+    title: str | None = Field(default=None, min_length=1, max_length=500)
+    status: str | None = Field(default=None, max_length=64)
+    severity: str | None = Field(default=None, max_length=32)
+
+
+def register_events(bus: EventBus, pool: asyncpg.Pool | None = None) -> None:
+    if pool is None:
+        return
+
+    async def on_alert(ev: DomainEvent) -> None:
+        if not isinstance(ev, AlertReceived) or ev.raw_payload_ref is None:
+            return
+        a = ev.alert
+        title = (a.title if a else "Алерт без названия")[:500]
+        sev = (a.severity.value if a else "warning")
+        try:
+            async with pool.acquire() as conn:
+                await conn.execute(
+                    """
+                    INSERT INTO incidents (
+                        title, status, severity, source, ingress_event_id,
+                        grafana_org_slug, service_name
+                    )
+                    VALUES ($1, 'open', $2, 'grafana', $3::uuid, $4, $5)
+                    """,
+                    title,
+                    sev,
+                    ev.raw_payload_ref,
+                    ev.grafana_org_slug,
+                    ev.service_name,
+                )
+        except Exception:
+            log.exception("incidents: не удалось создать инцидент из alert.received")
+
+    bus.subscribe("alert.received", on_alert)
+
+
+async def render_home_fragment(request: Request) -> str:
+    pool = get_pool(request)
+    if pool is None:
+        return '<p class="module-note">Нужна БД для списка инцидентов.</p>'
+    try:
+        async with pool.acquire() as conn:
+            n = await conn.fetchval("SELECT count(*)::int FROM incidents")
+    except Exception:
+        return '<p class="module-note">Таблица инцидентов недоступна (миграции?).</p>'
+    return f'<div class="module-fragment"><p>Инцидентов в учёте: <strong>{int(n)}</strong></p></div>'
+
+
+@router.get("/")
+async def list_incidents_api(
+    pool: asyncpg.Pool | None = Depends(get_pool),
+    limit: int = 50,
+    grafana_org_slug: str | None = None,
+    service_name: str | None = None,
+):
+    if pool is None:
+        return {"items": [], "database": "disabled"}
+    limit = min(max(limit, 1), 200)
+    conditions: list[str] = []
+    args: list = []
+    if grafana_org_slug and grafana_org_slug.strip():
+        args.append(grafana_org_slug.strip())
+        conditions.append(f"grafana_org_slug = ${len(args)}")
+    if service_name and service_name.strip():
+        args.append(service_name.strip())
+        conditions.append(f"service_name = ${len(args)}")
+    where_sql = ("WHERE " + " AND ".join(conditions)) if conditions else ""
+    args.append(limit)
+    lim_ph = f"${len(args)}"
+    q = f"""
+            SELECT id, title, status, severity, source, ingress_event_id, created_at, updated_at,
+                   grafana_org_slug, service_name
+            FROM incidents
+            {where_sql}
+            ORDER BY created_at DESC
+            LIMIT {lim_ph}
+            """
+    async with pool.acquire() as conn:
+        rows = await conn.fetch(q, *args)
+    items = []
+    for r in rows:
+        items.append(
+            {
+                "id": str(r["id"]),
+                "title": r["title"],
+                "status": r["status"],
+                "severity": r["severity"],
+                "source": r["source"],
+                "ingress_event_id": str(r["ingress_event_id"]) if r["ingress_event_id"] else None,
+                "created_at": r["created_at"].isoformat() if r["created_at"] else None,
+                "updated_at": r["updated_at"].isoformat() if r.get("updated_at") else None,
+                "grafana_org_slug": r.get("grafana_org_slug"),
+                "service_name": r.get("service_name"),
+            }
+        )
+    return {"items": items}
+
+
+@router.post("/", status_code=201)
+async def create_incident_api(
+    body: IncidentCreate,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            INSERT INTO incidents (title, status, severity, source, grafana_org_slug, service_name)
+            VALUES ($1, $2, $3, 'manual', NULL, NULL)
+            RETURNING id, title, status, severity, source, ingress_event_id, created_at, updated_at,
+                      grafana_org_slug, service_name
+            """,
+            body.title.strip(),
+            body.status,
+            body.severity,
+        )
+    return {
+        "id": str(row["id"]),
+        "title": row["title"],
+        "status": row["status"],
+        "severity": row["severity"],
+        "source": row["source"],
+        "ingress_event_id": None,
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+        "updated_at": row["updated_at"].isoformat() if row.get("updated_at") else None,
+        "grafana_org_slug": row.get("grafana_org_slug"),
+        "service_name": row.get("service_name"),
+    }
+
+
+def _incident_row_dict(row) -> dict:
+    return {
+        "id": str(row["id"]),
+        "title": row["title"],
+        "status": row["status"],
+        "severity": row["severity"],
+        "source": row["source"],
+        "ingress_event_id": str(row["ingress_event_id"]) if row["ingress_event_id"] else None,
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+        "updated_at": row["updated_at"].isoformat() if row.get("updated_at") else None,
+        "grafana_org_slug": row.get("grafana_org_slug"),
+        "service_name": row.get("service_name"),
+    }
+
+
+@router.get("/{incident_id}/tasks")
+async def list_incident_tasks_api(
+    incident_id: UUID,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+    limit: int = 100,
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    limit = min(max(limit, 1), 200)
+    async with pool.acquire() as conn:
+        exists = await conn.fetchval("SELECT 1 FROM incidents WHERE id = $1::uuid", incident_id)
+        if not exists:
+            raise HTTPException(status_code=404, detail="incident not found")
+        rows = await conn.fetch(
+            """
+            SELECT id, incident_id, title, status, created_at
+            FROM tasks WHERE incident_id = $1::uuid
+            ORDER BY created_at DESC LIMIT $2
+            """,
+            incident_id,
+            limit,
+        )
+    items = []
+    for r in rows:
+        items.append(
+            {
+                "id": str(r["id"]),
+                "incident_id": str(r["incident_id"]) if r["incident_id"] else None,
+                "title": r["title"],
+                "status": r["status"],
+                "created_at": r["created_at"].isoformat() if r["created_at"] else None,
+            }
+        )
+    return {"incident_id": str(incident_id), "items": items}
+
+
+@router.get("/{incident_id}")
+async def get_incident_api(incident_id: UUID, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            SELECT id, title, status, severity, source, ingress_event_id, created_at, updated_at,
+                   grafana_org_slug, service_name
+            FROM incidents WHERE id = $1::uuid
+            """,
+            incident_id,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return _incident_row_dict(row)
+
+
+@router.patch("/{incident_id}")
+async def patch_incident_api(
+    incident_id: UUID,
+    body: IncidentPatch,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    if body.title is None and body.status is None and body.severity is None:
+        raise HTTPException(status_code=400, detail="no fields to update")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            UPDATE incidents SET
+                title = COALESCE($2, title),
+                status = COALESCE($3, status),
+                severity = COALESCE($4, severity),
+                updated_at = now()
+            WHERE id = $1::uuid
+            RETURNING id, title, status, severity, source, ingress_event_id, created_at, updated_at,
+                      grafana_org_slug, service_name
+            """,
+            incident_id,
+            body.title.strip() if body.title is not None else None,
+            body.status,
+            body.severity,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return _incident_row_dict(row)
+
+
+@ui_router.get("/", response_class=HTMLResponse)
+async def incidents_ui_home(request: Request):
+    pool = get_pool(request)
+    rows_html = ""
+    err = ""
+    if pool is None:
+        err = "<p>База данных не настроена.</p>"
+    else:
+        try:
+            async with pool.acquire() as conn:
+                rows = await conn.fetch(
+                    """
+                    SELECT id, title, status, severity, source, created_at, grafana_org_slug, service_name
+                    FROM incidents
+                    ORDER BY created_at DESC
+                    LIMIT 100
+                    """
+                )
+            for r in rows:
+                org = html.escape(str(r["grafana_org_slug"] or "—"))
+                svc = html.escape(str(r["service_name"] or "—"))
+                rows_html += (
+                    "<tr>"
+                    f"<td>{html.escape(str(r['id']))[:8]}…</td>"
+                    f"<td>{html.escape(r['title'])}</td>"
+                    f"<td>{html.escape(r['status'])}</td>"
+                    f"<td>{html.escape(r['severity'])}</td>"
+                    f"<td>{html.escape(r['source'])}</td>"
+                    f"<td>{org}</td>"
+                    f"<td>{svc}</td>"
+                    "</tr>"
+                )
+        except Exception as e:
+            err = f"<p class=\"module-err\">{html.escape(str(e))}</p>"
+    inner = f"""<h1>Инциденты</h1>
+{err}
+<table class="irm-table">
+<thead><tr><th>ID</th><th>Заголовок</th><th>Статус</th><th>Важность</th><th>Источник</th><th>Grafana slug</th><th>Сервис</th></tr></thead>
+<tbody>{rows_html or '<tr><td colspan="7">Пока нет записей</td></tr>'}</tbody>
+</table>
+<p><small>Создание из Grafana: webhook → запись в <code>ingress_events</code> → событие → строка здесь.</small></p>"""
+    return HTMLResponse(
+        wrap_module_html_page(
+            document_title="Инциденты — onGuard24",
+            current_slug="incidents",
+            main_inner_html=inner,
+        )
+    )

onguard24/modules/registry.py

@@ -5,6 +5,7 @@
 
 from __future__ import annotations
 
+import asyncpg
 from collections.abc import Awaitable, Callable
 from dataclasses import dataclass
 
@@ -12,10 +13,19 @@ from fastapi import APIRouter
 from starlette.requests import Request
 
 from onguard24.domain.events import EventBus
-from onguard24.modules import contacts, schedules, statusboard
+from onguard24.modules import (
+    contacts,
+    escalations,
+    grafana_catalog,
+    incidents,
+    schedules,
+    statusboard,
+    tasks,
+)
 
 # async (Request) -> str — фрагмент HTML для главной страницы (опционально)
 HomeFragment = Callable[[Request], Awaitable[str]]
+RegisterEvents = Callable[[EventBus, asyncpg.Pool | None], None]
 
 
 @dataclass(frozen=True)
@@ -24,7 +34,7 @@ class ModuleMount:
 
     router: APIRouter
     url_prefix: str
-    register_events: Callable[[EventBus], None]
+    register_events: RegisterEvents
     slug: str
     title: str
     ui_router: APIRouter | None = None
@@ -33,6 +43,42 @@ class ModuleMount:
 
 def _mounts() -> list[ModuleMount]:
     return [
+        ModuleMount(
+            router=grafana_catalog.router,
+            url_prefix="/api/v1/modules/grafana-catalog",
+            register_events=grafana_catalog.register_events,
+            slug="grafana-catalog",
+            title="Каталог Grafana",
+            ui_router=grafana_catalog.ui_router,
+            render_home_fragment=grafana_catalog.render_home_fragment,
+        ),
+        ModuleMount(
+            router=incidents.router,
+            url_prefix="/api/v1/modules/incidents",
+            register_events=incidents.register_events,
+            slug="incidents",
+            title="Инциденты",
+            ui_router=incidents.ui_router,
+            render_home_fragment=incidents.render_home_fragment,
+        ),
+        ModuleMount(
+            router=tasks.router,
+            url_prefix="/api/v1/modules/tasks",
+            register_events=tasks.register_events,
+            slug="tasks",
+            title="Задачи",
+            ui_router=tasks.ui_router,
+            render_home_fragment=tasks.render_home_fragment,
+        ),
+        ModuleMount(
+            router=escalations.router,
+            url_prefix="/api/v1/modules/escalations",
+            register_events=escalations.register_events,
+            slug="escalations",
+            title="Эскалации",
+            ui_router=escalations.ui_router,
+            render_home_fragment=escalations.render_home_fragment,
+        ),
         ModuleMount(
             router=schedules.router,
             url_prefix="/api/v1/modules/schedules",
@@ -66,6 +112,6 @@ def _mounts() -> list[ModuleMount]:
 MODULE_MOUNTS: list[ModuleMount] = _mounts()
 
 
-def register_module_events(bus: EventBus) -> None:
+def register_module_events(bus: EventBus, pool: asyncpg.Pool | None = None) -> None:
     for m in MODULE_MOUNTS:
-        m.register_events(bus)
+        m.register_events(bus, pool)

onguard24/modules/schedules.py

@@ -9,7 +9,7 @@ router = APIRouter(tags=["module-schedules"])
 ui_router = APIRouter(tags=["web-schedules"], include_in_schema=False)
 
 
-def register_events(_bus: EventBus) -> None:
+def register_events(_bus: EventBus, _pool=None) -> None:
     """Подписка на доменные события (например alert.received)."""
     # _bus.subscribe("alert.received", handler)
 

onguard24/modules/statusboard.py

@@ -9,7 +9,7 @@ router = APIRouter(tags=["module-statusboard"])
 ui_router = APIRouter(tags=["web-statusboard"], include_in_schema=False)
 
 
-def register_events(_bus: EventBus) -> None:
+def register_events(_bus: EventBus, _pool=None) -> None:
     pass
 
 

onguard24/modules/tasks.py

@@ -0,0 +1,221 @@
+"""IRM: задачи по инцидентам (или вне привязки)."""
+
+from __future__ import annotations
+
+import html
+from uuid import UUID
+
+import asyncpg
+from fastapi import APIRouter, Depends, HTTPException, Request
+from fastapi.responses import HTMLResponse
+from pydantic import BaseModel, Field
+
+from onguard24.deps import get_pool
+from onguard24.domain.events import EventBus
+from onguard24.modules.ui_support import wrap_module_html_page
+
+router = APIRouter(tags=["module-tasks"])
+ui_router = APIRouter(tags=["web-tasks"], include_in_schema=False)
+
+
+class TaskCreate(BaseModel):
+    title: str = Field(..., min_length=1, max_length=500)
+    incident_id: UUID | None = None
+
+
+class TaskPatch(BaseModel):
+    title: str | None = Field(default=None, min_length=1, max_length=500)
+    status: str | None = Field(default=None, max_length=64)
+
+
+def register_events(_bus: EventBus, _pool: asyncpg.Pool | None = None) -> None:
+    pass
+
+
+async def render_home_fragment(request: Request) -> str:
+    pool = get_pool(request)
+    if pool is None:
+        return '<p class="module-note">Нужна БД для задач.</p>'
+    try:
+        async with pool.acquire() as conn:
+            n = await conn.fetchval("SELECT count(*)::int FROM tasks")
+    except Exception:
+        return '<p class="module-note">Таблица задач недоступна (миграции?).</p>'
+    return f'<div class="module-fragment"><p>Задач: <strong>{int(n)}</strong></p></div>'
+
+
+@router.get("/")
+async def list_tasks_api(
+    pool: asyncpg.Pool | None = Depends(get_pool),
+    incident_id: UUID | None = None,
+    limit: int = 100,
+):
+    if pool is None:
+        return {"items": [], "database": "disabled"}
+    limit = min(max(limit, 1), 200)
+    async with pool.acquire() as conn:
+        if incident_id:
+            rows = await conn.fetch(
+                """
+                SELECT id, incident_id, title, status, created_at
+                FROM tasks WHERE incident_id = $1::uuid
+                ORDER BY created_at DESC LIMIT $2
+                """,
+                incident_id,
+                limit,
+            )
+        else:
+            rows = await conn.fetch(
+                """
+                SELECT id, incident_id, title, status, created_at
+                FROM tasks
+                ORDER BY created_at DESC
+                LIMIT $1
+                """,
+                limit,
+            )
+    items = []
+    for r in rows:
+        items.append(
+            {
+                "id": str(r["id"]),
+                "incident_id": str(r["incident_id"]) if r["incident_id"] else None,
+                "title": r["title"],
+                "status": r["status"],
+                "created_at": r["created_at"].isoformat() if r["created_at"] else None,
+            }
+        )
+    return {"items": items}
+
+
+@router.post("/", status_code=201)
+async def create_task_api(body: TaskCreate, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    if body.incident_id:
+        async with pool.acquire() as conn:
+            ok = await conn.fetchval(
+                "SELECT 1 FROM incidents WHERE id = $1::uuid",
+                body.incident_id,
+            )
+        if not ok:
+            raise HTTPException(status_code=400, detail="incident not found")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            INSERT INTO tasks (title, incident_id, status)
+            VALUES ($1, $2::uuid, 'open')
+            RETURNING id, incident_id, title, status, created_at
+            """,
+            body.title.strip(),
+            body.incident_id,
+        )
+    return {
+        "id": str(row["id"]),
+        "incident_id": str(row["incident_id"]) if row["incident_id"] else None,
+        "title": row["title"],
+        "status": row["status"],
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+    }
+
+
+@router.get("/{task_id}")
+async def get_task_api(task_id: UUID, pool: asyncpg.Pool | None = Depends(get_pool)):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            SELECT id, incident_id, title, status, created_at
+            FROM tasks WHERE id = $1::uuid
+            """,
+            task_id,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return {
+        "id": str(row["id"]),
+        "incident_id": str(row["incident_id"]) if row["incident_id"] else None,
+        "title": row["title"],
+        "status": row["status"],
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+    }
+
+
+@router.patch("/{task_id}")
+async def patch_task_api(
+    task_id: UUID,
+    body: TaskPatch,
+    pool: asyncpg.Pool | None = Depends(get_pool),
+):
+    if pool is None:
+        raise HTTPException(status_code=503, detail="database disabled")
+    if body.title is None and body.status is None:
+        raise HTTPException(status_code=400, detail="no fields to update")
+    async with pool.acquire() as conn:
+        row = await conn.fetchrow(
+            """
+            UPDATE tasks SET
+                title = COALESCE($2, title),
+                status = COALESCE($3, status)
+            WHERE id = $1::uuid
+            RETURNING id, incident_id, title, status, created_at
+            """,
+            task_id,
+            body.title.strip() if body.title is not None else None,
+            body.status,
+        )
+    if not row:
+        raise HTTPException(status_code=404, detail="not found")
+    return {
+        "id": str(row["id"]),
+        "incident_id": str(row["incident_id"]) if row["incident_id"] else None,
+        "title": row["title"],
+        "status": row["status"],
+        "created_at": row["created_at"].isoformat() if row["created_at"] else None,
+    }
+
+
+@ui_router.get("/", response_class=HTMLResponse)
+async def tasks_ui_home(request: Request):
+    pool = get_pool(request)
+    rows_html = ""
+    err = ""
+    if pool is None:
+        err = "<p>База данных не настроена.</p>"
+    else:
+        try:
+            async with pool.acquire() as conn:
+                rows = await conn.fetch(
+                    """
+                    SELECT t.id, t.title, t.status, t.incident_id, t.created_at
+                    FROM tasks t
+                    ORDER BY t.created_at DESC
+                    LIMIT 100
+                    """
+                )
+            for r in rows:
+                iid = str(r["incident_id"])[:8] + "…" if r["incident_id"] else "—"
+                rows_html += (
+                    "<tr>"
+                    f"<td>{html.escape(str(r['id']))[:8]}…</td>"
+                    f"<td>{html.escape(r['title'])}</td>"
+                    f"<td>{html.escape(r['status'])}</td>"
+                    f"<td>{html.escape(iid)}</td>"
+                    "</tr>"
+                )
+        except Exception as e:
+            err = f"<p class=\"module-err\">{html.escape(str(e))}</p>"
+    inner = f"""<h1>Задачи</h1>
+{err}
+<table class="irm-table">
+<thead><tr><th>ID</th><th>Заголовок</th><th>Статус</th><th>Инцидент</th></tr></thead>
+<tbody>{rows_html or '<tr><td colspan="4">Пока нет задач</td></tr>'}</tbody>
+</table>"""
+    return HTMLResponse(
+        wrap_module_html_page(
+            document_title="Задачи — onGuard24",
+            current_slug="tasks",
+            main_inner_html=inner,
+        )
+    )

onguard24/modules/ui_support.py

@@ -27,6 +27,9 @@ APP_SHELL_CSS = """
     .rail-item a:hover { background: #f4f4f5; }
     .rail-item.is-active a { background: #eff6ff; color: #1d4ed8; font-weight: 600; }
     .module-page-main h1 { margin-top: 0; font-size: 1.35rem; }
+    .irm-table { width: 100%; border-collapse: collapse; font-size: 0.9rem; }
+    .irm-table th, .irm-table td { border: 1px solid #e4e4e7; padding: 0.45rem 0.65rem; text-align: left; }
+    .irm-table thead th { background: #f4f4f5; }
 """
 
 

onguard24/status_snapshot.py

@@ -3,6 +3,7 @@
 from fastapi import Request
 
 from onguard24.config import Settings
+from onguard24.grafana_sources import iter_grafana_sources
 from onguard24.integrations import forgejo_api, grafana_api
 from onguard24.vaultcheck import ping as vault_ping
 
@@ -31,31 +32,46 @@ async def build(request: Request) -> dict:
     else:
         out["vault"] = "disabled"
 
-    gu = settings.grafana_url.strip()
+    sources = iter_grafana_sources(settings)
-    if not gu:
+    if not sources:
         out["grafana"] = "disabled"
-    elif settings.grafana_service_account_token.strip():
+    else:
-        ok, err = await grafana_api.ping(gu, settings.grafana_service_account_token)
+        instances: list[dict] = []
+        for src in sources:
+            entry: dict = {"slug": src.slug, "url": src.api_url}
+            if src.api_token.strip():
+                ok, err = await grafana_api.ping(src.api_url, src.api_token)
                 if ok:
-            user, _ = await grafana_api.get_signed_in_user(gu, settings.grafana_service_account_token)
+                    user, _ = await grafana_api.get_signed_in_user(src.api_url, src.api_token)
-            entry: dict = {"status": "ok", "url": gu, "api": "authenticated"}
+                    entry["status"] = "ok"
+                    entry["api"] = "authenticated"
                     if user:
                         login = user.get("login") or user.get("email")
                         if login:
                             entry["service_account_login"] = login
-            out["grafana"] = entry
                 else:
-            out["grafana"] = {"status": "error", "detail": err, "url": gu}
+                    entry["status"] = "error"
+                    entry["detail"] = err
             else:
-        live_ok, live_err = await grafana_api.health_live(gu)
+                live_ok, live_err = await grafana_api.health_live(src.api_url)
                 if live_ok:
-            out["grafana"] = {
+                    entry["status"] = "reachable"
-                "status": "reachable",
+                    entry["detail"] = "задай api_token в GRAFANA_SOURCES_JSON для проверки API"
-                "url": gu,
-                "detail": "задай GRAFANA_SERVICE_ACCOUNT_TOKEN для вызовов API",
-            }
                 else:
-            out["grafana"] = {"status": "error", "detail": live_err, "url": gu}
+                    entry["status"] = "error"
+                    entry["detail"] = live_err
+            instances.append(entry)
+
+        if all(i.get("status") == "ok" for i in instances):
+            agg = "ok"
+        elif any(i.get("status") == "error" for i in instances):
+            agg = "error"
+        else:
+            agg = "reachable"
+        out["grafana"] = {
+            "status": agg,
+            "instances": instances,
+        }
 
     fj = settings.forgejo_url.strip()
     if not fj:

pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "onguard24"
-version = "1.4.0"
+version = "1.7.0"
 description = "onGuard24 — модульный сервис (аналог IRM)"
 readme = "README.md"
 requires-python = ">=3.11"

tests/conftest.py

@@ -14,6 +14,7 @@ for key in (
     "FORGEJO_URL",
     "FORGEJO_TOKEN",
     "GRAFANA_WEBHOOK_SECRET",
+    "GRAFANA_SOURCES_JSON",
 ):
     os.environ.pop(key, None)
 os.environ["DATABASE_URL"] = ""

tests/irm_db_fake.py

@@ -0,0 +1,286 @@
+"""In-memory «пул» для тестов IRM без реального PostgreSQL."""
+
+from __future__ import annotations
+
+from dataclasses import dataclass, field
+from datetime import datetime, timezone
+from typing import Any
+from uuid import UUID, uuid4
+
+
+def _now() -> datetime:
+    return datetime.now(timezone.utc)
+
+
+@dataclass
+class Row:
+    """Минимальная обёртка под asyncpg.Record (доступ по ключу)."""
+
+    _data: dict[str, Any]
+
+    def __getitem__(self, key: str) -> Any:
+        return self._data[key]
+
+    def get(self, key: str, default: Any = None) -> Any:
+        return self._data.get(key, default)
+
+
+class IrmFakeConn:
+    def __init__(self, store: IrmFakeStore) -> None:
+        self.store = store
+
+    def _q(self, query: str) -> str:
+        return " ".join(query.split())
+
+    async def execute(self, query: str, *args: Any) -> str:
+        q = self._q(query)
+        if "INSERT INTO incidents" in q and "ingress_event_id" in q:
+            self.store.insert_incident_alert(
+                args[0], args[1], args[2], args[3], args[4]
+            )
+            return "INSERT 0 1"
+        raise AssertionError(f"execute not implemented: {q[:80]}")
+
+    async def fetchval(self, query: str, *args: Any) -> Any:
+        q = self._q(query)
+        if "count(*)" in q and "FROM incidents" in q and "escalation" not in q:
+            return len(self.store.incidents)
+        if "count(*)" in q and "FROM tasks" in q:
+            return len(self.store.tasks)
+        if "count(*)" in q and "escalation_policies" in q:
+            return sum(1 for p in self.store.policies.values() if p["enabled"])
+        if "SELECT 1 FROM incidents WHERE id" in q:
+            uid = args[0]
+            return 1 if uid in self.store.incidents else None
+        raise AssertionError(f"fetchval not implemented: {q[:100]}")
+
+    async def fetch(self, query: str, *args: Any) -> list[Row]:
+        q = self._q(query)
+        if "FROM incidents" in q and "ORDER BY created_at DESC" in q:
+            rows = sorted(self.store.incidents.values(), key=lambda x: x["created_at"], reverse=True)
+            if "grafana_org_slug =" in q and "service_name =" in q:
+                rows = [
+                    r
+                    for r in rows
+                    if r.get("grafana_org_slug") == args[0]
+                    and r.get("service_name") == args[1]
+                ]
+                lim = args[2]
+            elif "grafana_org_slug =" in q:
+                rows = [r for r in rows if r.get("grafana_org_slug") == args[0]]
+                lim = args[1]
+            elif "service_name =" in q:
+                rows = [r for r in rows if r.get("service_name") == args[0]]
+                lim = args[1]
+            else:
+                lim = args[0]
+            return [Row(dict(r)) for r in rows[:lim]]
+        if "FROM tasks" in q and "WHERE incident_id" in q and "ORDER BY" in q:
+            iid, lim = args[0], args[1]
+            match = [t for t in self.store.tasks.values() if t["incident_id"] == iid]
+            match.sort(key=lambda x: x["created_at"], reverse=True)
+            return [Row(dict(t)) for t in match[:lim]]
+        if "FROM tasks" in q and "WHERE incident_id" in q:
+            iid, lim = args[0], args[1]
+            match = [t for t in self.store.tasks.values() if t["incident_id"] == iid]
+            match.sort(key=lambda x: x["created_at"], reverse=True)
+            return [Row(dict(t)) for t in match[:lim]]
+        if "FROM tasks t" in q or ("FROM tasks" in q and "ORDER BY t.created_at" in q):
+            lim = args[0]
+            rows = sorted(self.store.tasks.values(), key=lambda x: x["created_at"], reverse=True)[:lim]
+            return [Row(dict(r)) for r in rows]
+        if "FROM tasks" in q and "ORDER BY created_at DESC" in q and "WHERE" not in q:
+            lim = args[0]
+            rows = sorted(self.store.tasks.values(), key=lambda x: x["created_at"], reverse=True)[:lim]
+            return [Row(dict(r)) for r in rows]
+        if "FROM escalation_policies" in q and "ORDER BY name" in q:
+            rows = sorted(self.store.policies.values(), key=lambda x: x["name"])
+            return [Row(dict(r)) for r in rows]
+        raise AssertionError(f"fetch not implemented: {q[:120]}")
+
+    async def fetchrow(self, query: str, *args: Any) -> Row | None:
+        q = self._q(query)
+        if "INSERT INTO incidents" in q and "VALUES ($1, $2, $3, 'manual'" in q:
+            return Row(self.store.insert_incident_manual(args[0], args[1], args[2]))
+        if "FROM incidents WHERE id" in q and "UPDATE" not in q and "/tasks" not in query.lower():
+            return self.store.get_incident(args[0])
+        if "UPDATE incidents SET" in q:
+            return self.store.update_incident(args[0], args[1], args[2], args[3])
+        if "INSERT INTO tasks" in q:
+            return Row(self.store.insert_task(args[0], args[1]))
+        if "FROM tasks WHERE id" in q and "UPDATE" not in q:
+            tid = args[0]
+            t = self.store.tasks.get(tid)
+            return Row(dict(t)) if t else None
+        if "UPDATE tasks SET" in q:
+            return self.store.update_task(args[0], args[1], args[2])
+        if "INSERT INTO escalation_policies" in q:
+            return Row(self.store.insert_policy(args[0], args[1], args[2]))
+        if "FROM escalation_policies WHERE id" in q and "UPDATE" not in q and "DELETE" not in q:
+            pid = args[0]
+            p = self.store.policies.get(pid)
+            return Row(dict(p)) if p else None
+        if "UPDATE escalation_policies SET" in q:
+            return self.store.update_policy(args[0], args[1], args[2], args[3])
+        if "DELETE FROM escalation_policies" in q:
+            return self.store.delete_policy(args[0])
+        raise AssertionError(f"fetchrow not implemented: {q[:120]}")
+
+
+@dataclass
+class IrmFakeStore:
+    incidents: dict[UUID, dict[str, Any]] = field(default_factory=dict)
+    tasks: dict[UUID, dict[str, Any]] = field(default_factory=dict)
+    policies: dict[UUID, dict[str, Any]] = field(default_factory=dict)
+
+    def insert_incident_alert(
+        self,
+        title: str,
+        sev: str,
+        ingress_id: UUID,
+        grafana_org_slug: Any,
+        service_name: Any,
+    ) -> None:
+        iid = uuid4()
+        now = _now()
+        self.incidents[iid] = {
+            "id": iid,
+            "title": title,
+            "status": "open",
+            "severity": sev,
+            "source": "grafana",
+            "ingress_event_id": ingress_id,
+            "created_at": now,
+            "updated_at": now,
+            "grafana_org_slug": grafana_org_slug,
+            "service_name": service_name,
+        }
+
+    def insert_incident_manual(self, title: str, status: str, severity: str) -> dict[str, Any]:
+        iid = uuid4()
+        now = _now()
+        row = {
+            "id": iid,
+            "title": title,
+            "status": status,
+            "severity": severity,
+            "source": "manual",
+            "ingress_event_id": None,
+            "created_at": now,
+            "updated_at": now,
+            "grafana_org_slug": None,
+            "service_name": None,
+        }
+        self.incidents[iid] = row
+        return row
+
+    def get_incident(self, iid: UUID) -> Row | None:
+        r = self.incidents.get(iid)
+        return Row(dict(r)) if r else None
+
+    def update_incident(
+        self,
+        iid: UUID,
+        title: str | None,
+        status: str | None,
+        severity: str | None,
+    ) -> Row | None:
+        r = self.incidents.get(iid)
+        if not r:
+            return None
+        if title is not None:
+            r["title"] = title
+        if status is not None:
+            r["status"] = status
+        if severity is not None:
+            r["severity"] = severity
+        r["updated_at"] = _now()
+        return Row(dict(r))
+
+    def insert_task(self, title: str, incident_id: UUID | None) -> dict[str, Any]:
+        tid = uuid4()
+        now = _now()
+        row = {
+            "id": tid,
+            "incident_id": incident_id,
+            "title": title,
+            "status": "open",
+            "created_at": now,
+        }
+        self.tasks[tid] = row
+        return row
+
+    def update_task(self, tid: UUID, title: str | None, status: str | None) -> Row | None:
+        r = self.tasks.get(tid)
+        if not r:
+            return None
+        if title is not None:
+            r["title"] = title
+        if status is not None:
+            r["status"] = status
+        return Row(dict(r))
+
+    def insert_policy(self, name: str, enabled: bool, steps_json: str) -> dict[str, Any]:
+        import json
+
+        pid = uuid4()
+        now = _now()
+        steps = json.loads(steps_json)
+        row = {
+            "id": pid,
+            "name": name,
+            "enabled": enabled,
+            "steps": steps,
+            "created_at": now,
+        }
+        self.policies[pid] = row
+        return row
+
+    def update_policy(
+        self,
+        pid: UUID,
+        name: str | None,
+        enabled: bool | None,
+        steps_json: str | None,
+    ) -> Row | None:
+        import json
+
+        r = self.policies.get(pid)
+        if not r:
+            return None
+        if name is not None:
+            r["name"] = name
+        if enabled is not None:
+            r["enabled"] = enabled
+        if steps_json is not None:
+            r["steps"] = json.loads(steps_json)
+        return Row(dict(r))
+
+    def delete_policy(self, pid: UUID) -> Row | None:
+        if pid not in self.policies:
+            return None
+        self.policies.pop(pid)
+        return Row({"id": pid})
+
+
+class IrmFakeAcquire:
+    def __init__(self, store: IrmFakeStore) -> None:
+        self.store = store
+
+    async def __aenter__(self) -> IrmFakeConn:
+        return IrmFakeConn(self.store)
+
+    async def __aexit__(self, *args: Any) -> None:
+        pass
+
+
+class IrmFakePool:
+    def __init__(self, store: IrmFakeStore | None = None) -> None:
+        self._store = store or IrmFakeStore()
+
+    def acquire(self) -> IrmFakeAcquire:
+        return IrmFakeAcquire(self._store)
+
+    @property
+    def store(self) -> IrmFakeStore:
+        return self._store

tests/test_grafana_catalog_api.py

@@ -0,0 +1,19 @@
+"""HTTP-обёртки каталога Grafana без реальной БД."""
+
+from fastapi.testclient import TestClient
+
+
+def test_grafana_catalog_sync_requires_db(client: TestClient) -> None:
+    r = client.post("/api/v1/modules/grafana-catalog/sync", json={})
+    assert r.status_code == 503
+
+
+def test_grafana_catalog_meta_no_db(client: TestClient) -> None:
+    r = client.get("/api/v1/modules/grafana-catalog/meta")
+    assert r.status_code == 200
+    assert r.json().get("database") == "disabled"
+
+
+def test_grafana_catalog_tree_requires_db(client: TestClient) -> None:
+    r = client.get("/api/v1/modules/grafana-catalog/tree?instance_slug=default")
+    assert r.status_code == 503

tests/test_grafana_topology.py

@@ -0,0 +1,93 @@
+"""Парсинг Ruler / слияние папок и HTTP-mock синхронизации Grafana."""
+
+import httpx
+import pytest
+import respx
+
+from onguard24.integrations.grafana_topology import (
+    merge_folder_rows,
+    parse_ruler_rules,
+)
+from onguard24.modules.grafana_catalog import pull_topology
+
+
+def test_parse_ruler_grafana_managed() -> None:
+    data = {
+        "nginx": [
+            {
+                "name": "prometheus (Nginx)",
+                "interval": "60s",
+                "rules": [
+                    {
+                        "grafana_alert": {"uid": "uid1", "title": "Nginx Down"},
+                        "labels": {"service": "nginx", "severity": "critical"},
+                    }
+                ],
+            }
+        ]
+    }
+    rows = parse_ruler_rules(data)
+    assert len(rows) == 1
+    assert rows[0].namespace_uid == "nginx"
+    assert rows[0].rule_uid == "uid1"
+    assert rows[0].rule_group_name == "prometheus (Nginx)"
+    assert rows[0].labels["service"] == "nginx"
+
+
+def test_parse_ruler_skips_non_grafana_alert() -> None:
+    data = {"x": [{"name": "g", "rules": [{"expr": "1"}]}]}
+    assert parse_ruler_rules(data) == []
+
+
+def test_merge_folder_rows_adds_namespaces() -> None:
+    api = [{"uid": "system", "title": "System", "parentUid": None}]
+    merged = merge_folder_rows(api, {"nginx", "system"})
+    uids = {m[0] for m in merged}
+    assert uids == {"system", "nginx"}
+    titles = {m[0]: m[1] for m in merged}
+    assert titles["nginx"] == "nginx"
+
+
+@pytest.mark.asyncio
+@respx.mock
+async def test_pull_topology_end_to_end() -> None:
+    base = "https://grafana.example.com"
+    respx.get(f"{base}/api/org").mock(
+        return_value=httpx.Response(200, json={"id": 3, "name": "adibrov"})
+    )
+    def _folders(request: httpx.Request) -> httpx.Response:
+        if "parentUid" in str(request.url):
+            return httpx.Response(200, json=[])
+        return httpx.Response(
+            200,
+            json=[{"uid": "nginx", "title": "Nginx Alerts", "parentUid": None}],
+        )
+
+    respx.get(f"{base}/api/folders").mock(side_effect=_folders)
+    ruler_body = {
+        "nginx": [
+            {
+                "name": "grp",
+                "interval": "1m",
+                "rules": [
+                    {
+                        "grafana_alert": {"uid": "r1", "title": "Down"},
+                        "labels": {"service": "nginx"},
+                    }
+                ],
+            }
+        ]
+    }
+    respx.get(f"{base}/api/ruler/grafana/api/v1/rules").mock(
+        return_value=httpx.Response(200, json=ruler_body)
+    )
+
+    out, err = await pull_topology(base, "test-token")
+    assert err is None
+    assert out is not None
+    assert out.org_id == 3
+    assert out.org_name == "adibrov"
+    assert len(out.folder_rows) == 1
+    assert out.folder_rows[0][0] == "nginx"
+    assert len(out.rules) == 1
+    assert out.rules[0].rule_uid == "r1"

tests/test_ingress.py

@@ -63,6 +63,35 @@ def test_grafana_webhook_inserts_with_mock_pool(client: TestClient) -> None:
         app.state.pool = real_pool
 
 
+def test_grafana_webhook_auto_org_from_external_url(client: TestClient) -> None:
+    from uuid import uuid4
+
+    mock_conn = AsyncMock()
+    uid = uuid4()
+    mock_conn.fetchrow = AsyncMock(return_value={"id": uid})
+    mock_cm = AsyncMock()
+    mock_cm.__aenter__ = AsyncMock(return_value=mock_conn)
+    mock_cm.__aexit__ = AsyncMock(return_value=None)
+    mock_pool = MagicMock()
+    mock_pool.acquire = MagicMock(return_value=mock_cm)
+
+    app = client.app
+    real_pool = app.state.pool
+    app.state.pool = mock_pool
+    try:
+        r = client.post(
+            "/api/v1/ingress/grafana",
+            content=json.dumps(
+                {"externalURL": "https://grafana-adibrov.example.com/", "title": "x"}
+            ),
+            headers={"Content-Type": "application/json"},
+        )
+        assert r.status_code == 202
+        assert mock_conn.fetchrow.call_args[0][3] == "grafana-adibrov.example.com"
+    finally:
+        app.state.pool = real_pool
+
+
 def test_grafana_webhook_publishes_alert_received(client: TestClient) -> None:
     from unittest.mock import patch
     from uuid import uuid4
@@ -92,3 +121,66 @@ def test_grafana_webhook_publishes_alert_received(client: TestClient) -> None:
             assert spy.await_args.kwargs.get("raw_payload_ref") == uid
         finally:
             app.state.pool = real_pool
+
+
+def test_grafana_webhook_org_any_slug_without_json_config(client: TestClient) -> None:
+    """Путь /{slug} не требует GRAFANA_SOURCES_JSON — slug просто сохраняется."""
+    from uuid import uuid4
+
+    mock_conn = AsyncMock()
+    uid = uuid4()
+    mock_conn.fetchrow = AsyncMock(return_value={"id": uid})
+    mock_cm = AsyncMock()
+    mock_cm.__aenter__ = AsyncMock(return_value=mock_conn)
+    mock_cm.__aexit__ = AsyncMock(return_value=None)
+    mock_pool = MagicMock()
+    mock_pool.acquire = MagicMock(return_value=mock_cm)
+
+    app = client.app
+    real_pool = app.state.pool
+    app.state.pool = mock_pool
+    try:
+        r = client.post(
+            "/api/v1/ingress/grafana/other",
+            content=b"{}",
+            headers={"Content-Type": "application/json"},
+        )
+        assert r.status_code == 202
+        assert mock_conn.fetchrow.call_args[0][3] == "other"
+    finally:
+        app.state.pool = real_pool
+
+
+def test_grafana_webhook_org_ok(client: TestClient) -> None:
+    from uuid import uuid4
+
+    mock_conn = AsyncMock()
+    uid = uuid4()
+    mock_conn.fetchrow = AsyncMock(return_value={"id": uid})
+    mock_cm = AsyncMock()
+    mock_cm.__aenter__ = AsyncMock(return_value=mock_conn)
+    mock_cm.__aexit__ = AsyncMock(return_value=None)
+    mock_pool = MagicMock()
+    mock_pool.acquire = MagicMock(return_value=mock_cm)
+
+    app = client.app
+    real_json = app.state.settings.grafana_sources_json
+    real_pool = app.state.pool
+    app.state.settings.grafana_sources_json = (
+        '[{"slug":"adibrov","api_url":"http://192.168.0.1:3000","api_token":"","webhook_secret":""}]'
+    )
+    app.state.pool = mock_pool
+    try:
+        r = client.post(
+            "/api/v1/ingress/grafana/adibrov",
+            content='{"title":"t"}',
+            headers={"Content-Type": "application/json"},
+        )
+        assert r.status_code == 202
+        call = mock_conn.fetchrow.call_args
+        assert "org_slug" in call[0][0].lower() or "org_slug" in str(call)
+        assert call[0][1] == "grafana"
+        assert call[0][3] == "adibrov"
+    finally:
+        app.state.settings.grafana_sources_json = real_json
+        app.state.pool = real_pool

tests/test_irm_api_with_fake_db.py

@@ -0,0 +1,118 @@
+"""IRM API с подменённым пулом БД (без PostgreSQL)."""
+
+from __future__ import annotations
+
+import pytest
+from fastapi.testclient import TestClient
+from starlette.requests import Request
+
+from onguard24.deps import get_pool
+from onguard24.main import app
+
+from tests.irm_db_fake import IrmFakePool
+
+
+@pytest.fixture
+def irm_client() -> tuple[TestClient, IrmFakePool]:
+    pool = IrmFakePool()
+
+    def override_get_pool(_request: Request):
+        return pool
+
+    app.dependency_overrides[get_pool] = override_get_pool
+    with TestClient(app) as client:
+        yield client, pool
+    app.dependency_overrides.pop(get_pool, None)
+
+
+def test_irm_incident_crud_and_tasks(irm_client: tuple[TestClient, IrmFakePool]) -> None:
+    client, _pool = irm_client
+    r = client.post(
+        "/api/v1/modules/incidents/",
+        json={"title": "Сбой API", "status": "open", "severity": "critical"},
+    )
+    assert r.status_code == 201
+    iid = r.json()["id"]
+    assert r.json()["source"] == "manual"
+
+    r = client.get(f"/api/v1/modules/incidents/{iid}")
+    assert r.status_code == 200
+    assert r.json()["title"] == "Сбой API"
+
+    r = client.patch(f"/api/v1/modules/incidents/{iid}", json={"status": "resolved"})
+    assert r.status_code == 200
+    assert r.json()["status"] == "resolved"
+
+    r = client.post(
+        "/api/v1/modules/tasks/",
+        json={"title": "Разбор логов", "incident_id": iid},
+    )
+    assert r.status_code == 201
+    tid = r.json()["id"]
+
+    r = client.get(f"/api/v1/modules/incidents/{iid}/tasks")
+    assert r.status_code == 200
+    assert len(r.json()["items"]) == 1
+    assert r.json()["items"][0]["id"] == tid
+
+    r = client.get(f"/api/v1/modules/tasks/{tid}")
+    assert r.status_code == 200
+    assert r.json()["status"] == "open"
+
+    r = client.patch(f"/api/v1/modules/tasks/{tid}", json={"status": "done"})
+    assert r.status_code == 200
+    assert r.json()["status"] == "done"
+
+
+def test_irm_task_bad_incident(irm_client: tuple[TestClient, IrmFakePool]) -> None:
+    client, _ = irm_client
+    import uuid
+
+    r = client.post(
+        "/api/v1/modules/tasks/",
+        json={"title": "x", "incident_id": str(uuid.uuid4())},
+    )
+    assert r.status_code == 400
+    assert r.json()["detail"] == "incident not found"
+
+
+def test_irm_incident_tasks_unknown(irm_client: tuple[TestClient, IrmFakePool]) -> None:
+    client, _ = irm_client
+    import uuid
+
+    rid = str(uuid.uuid4())
+    r = client.get(f"/api/v1/modules/incidents/{rid}/tasks")
+    assert r.status_code == 404
+
+
+def test_irm_patch_validation(irm_client: tuple[TestClient, IrmFakePool]) -> None:
+    client, _ = irm_client
+    r = client.post("/api/v1/modules/incidents/", json={"title": "t"})
+    iid = r.json()["id"]
+    r = client.patch(f"/api/v1/modules/incidents/{iid}", json={})
+    assert r.status_code == 400
+
+
+def test_irm_escalations_crud(irm_client: tuple[TestClient, IrmFakePool]) -> None:
+    client, _ = irm_client
+    r = client.post(
+        "/api/v1/modules/escalations/",
+        json={"name": "L1", "enabled": True, "steps": [{"after_min": 5, "channel": "slack"}]},
+    )
+    assert r.status_code == 201
+    pid = r.json()["id"]
+    assert r.json()["steps"][0]["channel"] == "slack"
+
+    r = client.get(f"/api/v1/modules/escalations/{pid}")
+    assert r.status_code == 200
+    assert r.json()["name"] == "L1"
+
+    r = client.patch(f"/api/v1/modules/escalations/{pid}", json={"enabled": False})
+    assert r.status_code == 200
+    assert r.json()["enabled"] is False
+
+    r = client.delete(f"/api/v1/modules/escalations/{pid}")
+    assert r.status_code == 204
+
+    r = client.get(f"/api/v1/modules/escalations/{pid}")
+    assert r.status_code == 404

tests/test_irm_modules.py

@@ -0,0 +1,71 @@
+"""IRM-модули: API без БД и обработчик инцидента по событию."""
+
+from unittest.mock import AsyncMock, MagicMock
+from uuid import uuid4
+
+import pytest
+from fastapi.testclient import TestClient
+
+from onguard24.domain.entities import Alert, Severity
+from onguard24.domain.events import AlertReceived
+
+
+def test_incidents_api_list_no_db(client: TestClient) -> None:
+    r = client.get("/api/v1/modules/incidents/")
+    assert r.status_code == 200
+    assert r.json() == {"items": [], "database": "disabled"}
+
+
+def test_tasks_api_list_no_db(client: TestClient) -> None:
+    r = client.get("/api/v1/modules/tasks/")
+    assert r.status_code == 200
+    assert r.json()["database"] == "disabled"
+
+
+def test_escalations_api_list_no_db(client: TestClient) -> None:
+    r = client.get("/api/v1/modules/escalations/")
+    assert r.status_code == 200
+    assert r.json()["database"] == "disabled"
+
+
+@pytest.mark.asyncio
+async def test_incident_inserted_on_alert_received() -> None:
+    """При пуле БД подписка создаёт инцидент (INSERT)."""
+    inserted: dict = {}
+
+    async def fake_execute(_query, *args):
+        inserted["args"] = args
+        return "INSERT 0 1"
+
+    mock_conn = AsyncMock()
+    mock_conn.execute = fake_execute
+    mock_cm = AsyncMock()
+    mock_cm.__aenter__ = AsyncMock(return_value=mock_conn)
+    mock_cm.__aexit__ = AsyncMock(return_value=None)
+    mock_pool = MagicMock()
+    mock_pool.acquire = MagicMock(return_value=mock_cm)
+
+    from onguard24.domain.events import InMemoryEventBus
+    from onguard24.modules import incidents as inc_mod
+
+    bus = InMemoryEventBus()
+    inc_mod.register_events(bus, mock_pool)
+
+    uid = uuid4()
+    ev = AlertReceived(
+        alert=Alert(source="grafana", title="CPU high", severity=Severity.WARNING),
+        raw_payload_ref=uid,
+    )
+    await bus.publish(ev)
+
+    assert inserted.get("args") is not None
+    assert inserted["args"][0] == "CPU high"
+    assert inserted["args"][1] == "warning"
+    assert inserted["args"][2] == uid
+    assert inserted["args"][3] is None
+    assert inserted["args"][4] is None
+
+
+def test_incidents_post_requires_db(client: TestClient) -> None:
+    r = client.post("/api/v1/modules/incidents/", json={"title": "x"})
+    assert r.status_code == 503

tests/test_root_ui.py

@@ -32,6 +32,10 @@ def test_rail_lists_all_registered_ui_modules(client: TestClient) -> None:
     assert r.status_code == 200
     t = r.text
     expected = (
+        ("grafana-catalog", "Каталог Grafana"),
+        ("incidents", "Инциденты"),
+        ("tasks", "Задачи"),
+        ("escalations", "Эскалации"),
         ("schedules", "Календарь дежурств"),
         ("contacts", "Контакты"),
         ("statusboard", "Светофор"),
@@ -43,7 +47,15 @@ def test_rail_lists_all_registered_ui_modules(client: TestClient) -> None:
 
 def test_each_module_page_single_active_nav_item(client: TestClient) -> None:
     """На странице модуля ровно один пункт с aria-current (текущий раздел)."""
-    for slug in ("schedules", "contacts", "statusboard"):
+    for slug in (
+        "grafana-catalog",
+        "incidents",
+        "tasks",
+        "escalations",
+        "schedules",
+        "contacts",
+        "statusboard",
+    ):
         r = client.get(f"/ui/modules/{slug}/")
         assert r.status_code == 200
         assert r.text.count('aria-current="page"') == 1

tests/test_status.py

@@ -42,6 +42,7 @@ def test_status_with_mocks(client: TestClient) -> None:
             vault_token="t",
             grafana_url="https://grafana.example",
             grafana_service_account_token="g",
+            grafana_sources_json="",
             forgejo_url="https://git.example",
             forgejo_token="f",
             grafana_webhook_secret="",
@@ -57,5 +58,6 @@ def test_status_with_mocks(client: TestClient) -> None:
         d = r.json()
         assert d["vault"]["status"] == "ok"
         assert d["grafana"]["status"] == "ok"
-        assert d["grafana"].get("service_account_login") == "tester"
+        assert len(d["grafana"]["instances"]) == 1
+        assert d["grafana"]["instances"][0].get("service_account_login") == "tester"
         assert d["forgejo"]["status"] == "ok"