kludex / fastapi-prometheus-grafana

FasAPI + Prometheus + Grafana! :tada:

FastAPI + Prometheus + Grafana 🎉

This is a minimal setup that you can build to monitor your FastAPI microservice.

Installation

There are only two prerequisites:

• Docker
• Docker-compose

Having both, you'll need to clone the repository:

git clone https://github.com/Kludex/fastapi-prometheus-grafana

Usage

You'll need to run the docker containers:

docker-compose up

Now you have access to those three containers and their respective ports:

• Prometheus: http://localhost:9090/
• Grafana: http://localhost:3000/
• FastAPI: http://localhost:8000/

On the FastAPI, you can access /metrics endpoint to see the data Prometheus is scraping from it.

How it looks like

References

• Prometheus FastAPI Instrumentator
• Generate and Track Metrics for Flask API Applications Using Prometheus and Grafana
• PromQL for Humans

fastapi-prometheus-grafana's Issues

Grafana 8.5.3 does not know dashboard

On my Ubuntu 20.04 running Docker 20.10.12 after docker-compose up --build I can see all 3 services up and running,
but Grafana does not show any fastAPI Dashboard.
Service app runs on 238.11 and shows HTTP 200 GET /metrics
Service prometheus seems to be ok as well,
only grafana does not come up with any dashboard.
I tried to change the target volume for
./datasource.yml:/etc/grafana/provisioning/datasource.yml to
/etc/grafana/provisioning/datasources/datasource.yml
but without luck. How does fastapi-dashboard.json get mounted into grafana?
Can anybody please give me a hint what's missing?

Errors per seconds graph

I have looked into the way in which the ErrorsPerSecond graph's PromQL qeury is constructed. I believe that instead of using the http_request_duration_seconds_count metric it should be using the http_requests_total metric instead.

Rework project

PRs are welcome to make this pretty, and/or more usable.