Observability is a cornerstone of the software development cycle, and we've made it easier to integrate Tinybird with your preferred observability tools.
Tinybird has released support for the Prometheus endpoint format on all published API Endpoints. In addition to generating JSON, CSV, NDJSON, and Parquet formats, you can now build endpoints specifically to be consumed in Prometheus format.
Simply use the following URL structure:
You can get more info on how to use the Prometheus endpoint format, including how to structure your query results to match the Prometheus data model, in the Prometheus endpoint docs.
Why Prometheus?
Tinybird is data infrastructure for software builders, and in most cases, software builders use JSON endpoints because that's standard for web application development.
But building the app is only part of the development lifecycle, and observability can't be overlooked. With the Prometheus endpoint format, you can monitor not only your Tinybird infrastructure (using free Service Data Sources), but also anything you have in Tinybird.
Prometheus is an open format for metrics adopted by all the top observability platforms. [Grafana](https://www.tinybird.co/blog-posts/clickhouse-grafana-example), Datadog, and many others integrate Prometheus endpoints for visualization and alerting, and by adapting our endpoint formats to include Prometheus, we've made it easier to build comprehensive observability systems in the monitoring platform you prefer.
Some examples
We've already built out a couple examples of some great observability projects using Prometheus endpoints.
Tinybird Organization Monitoring
Relevant to any Tinybird user with more than one Workspace, a project we call the "Organization Export" - a single Prometheus endpoint with all the metrics you need to keep tabs on your Tinybird Workspaces.
You can peep the repo for that project here (it even includes sample configs for Grafana and Datadog dashboards).
MCP Server Analytics
If you've had your ear to the ground on developments in AI, you're likely aware of Anthropic's open source Model Context Protocol project designed to provide an open server-client framework for providing context to LLM applications.
We set up a Tinybird MCP Server so we could converse with our Tinybird Workspaces via the Claude Desktop app, and quickly added logging functionality to monitor locally-deployed servers and aggregate logs metrics. We use the Prometheus format for endpoints to visualize MCP Server metrics in Grafana. Check out the repo for MCP server analytics, which includes another sample dashboard for Grafana.
What's next?
Tinybird has invested in great observability from day one. We include observability dashboards within our UI and allow you to query Service Data Sources for free to build metrics on your reads and writes to the database.
Follow along for more examples of monitoring and observability with Tinybird. If you're looking for practical advice on utilizing the Prometheus endpoint format in your o11y projects, join our Slack community.
