This is a test repository. I want to validate what is the promxy behavior when one of its targets is down.
We build two docker-composes that runs two Prometheus (cluster-a and cluster-b). They scrapes two sample apps (go-prometheus-app),
which just export some metrics. Moreover, we have one container running Promxy, which is configured to query both Prometheus.
The docker-compose-prometheus.yml run with the ignore_error flag as true and the docker-compose-prometheus-ignore-errors.yml as false.
Run the desired docker-compose. Ex.:
docker-compose -f docker-compose-prometheus-ignore-errors.yml up
Access the promxy UI at http://localhost:8082/ and run the query:
sum(static_counter)
This is a metric collected by our go-prometheus-app which always return 1, so as we have two Prometheus, this should return 2.