This integration supports linking sentry issues with a notion database. Written in python, it supplies a dockerfile that can run a microservice fastapi application which exposes a few endpoints for sentry to use to communicate with notion.
Issue tracking in Sentry
Users can create new Notion database entries directly from Sentry issues.
Users can link existing Notion database entries to Sentry issues.
- Redis: A Redis server is required for caching database metadata and user information.
- Sentry: A Sentry internal integration
- Notion: A Notion integration
Follow the directions for setting up a Sentry internal integration. Store the client secret, to be provided in the SENTRY_NOTION_INTEGRATION_CLIENT_SECRET environment variable.
Copy paste the contents of the sentry_ui_schema.json file into the Sentry UI schema editor within the Sentry app.
After your application is deployed, provide its URL under Webhook URL in the Sentry app.
Follow the directions for setting up a Notion integration. Store the integration token, to be provided in the NOTION_TOKEN environment variable.
Find your database id, to be provided in the NOTION_CONFIG environment variable. To find your database id, see Notion's api docs.
The integration assumes you have a Notion database set up with the following columns:
ID(type: unique_id)Assignee(type: people)Sentry Link(type: url)
If you have different names for your columns, you can provide them in the NOTION_CONFIG environment variable.
Note: You will need to give your notion integration access to your database. This is found in the database settings under "Connections".
Build the provided Dockerfile and push it to your container registry. Then, deploy the application to your preferred platform, providing the required environment variables and ensuring the application can connect to the redis server.
The integration requires the following environment variables to be set.
NOTION_TOKEN: The notion API integration token.NOTION_CONFIG: JSON configuration for the Notion database (see below)SENTRY_NOTION_INTEGRATION_CLIENT_SECRET: The client secret for validating Sentry requests.
The application also provides the following environment variables:
CACHE_TIMEOUT: The cache timeout in seconds (default: 21600)REDIS_HOST: The Redis host (default:redis)REDIS_PORT: The Redis port (default:6379)
The NOTION_CONFIG environment variable should contain a JSON object with the following structure:
{
"database_id": "your-notion-database-id",
"column_names": {
"assignee": "notion-people-property-name",
"id": "notion-unique-id-property-name",
"sentry_url": "notion-url-property-name"
}
}The NOTION_CONFIG environment variable is used to map between the Notion database columns and the properties we want to use in the integration. The property names should match the names of the properties in the Notion database. Changes to the property names in the Notion database will require updating the NOTION_CONFIG environment variable.
All requests from Sentry are authenticated using HMAC signature validation. The integration verifies that requests are coming from Sentry by checking the sentry-hook-signature or sentry-app-signature headers against the SENTRY_NOTION_INTEGRATION_CLIENT_SECRET. This secret is found on the Sentry integrations settings page.
The integration uses Redis to cache Notion database metadata and user information to improve performance and reduce API calls to Notion. The cache timeout is set to 6 hours by default but can be adjusted using the CACHE_TIMEOUT environment variable.
The sentry_ui_schema.json file defines the UI components that appear in the Sentry interface. Changes to this file need to be copy pasted into the Sentry UI schema editor within the Sentry app.
Set your environment variables in the .env file.
docker-compose up
uv run pytest
uv run mypy src
uv run pytest
To test sentry requests to your application while developing locally, you can use ngrok. Run an ngrok tunnel and modify the webhook URL in the Sentry app to use the ngrok URL.