The backend health monitor is another container in the wdqs pod (see T425007) which provides two functions to begin with:
1. On startup it compares the timestamp in the `wikidata_index_meta.json` file with the timestamp in the index itself (see [[ https://gerrit.wikimedia.org/r/plugins/gitiles/operations/cookbooks/+/refs/heads/master/cookbooks/sre/wdqs/data-reload.py#360 | data-reload.py ]] for how to query the timestamp)
2. Expose a startup endpoint (as in [[ https://kubernetes.io/docs/concepts/configuration/liveness-readiness-startup-probes/#startup-probe | Kubernetes startup probe ]]) which verifies the consumer update lag against a configured threshold. This will ultimately be used to verify whether the pod can be pooled and receive requests.
The health service is supposed to be implemented as a Quarkus Java application (similar to [[ https://gitlab.wikimedia.org/repos/wikidata-platform/wdqs/wdqs-proxy | wdqs-proxy ]]).
AC:
[] New Quarkus service which can read the update lag from the consumer (configurable endpoint)
[] The service compares timestamps and fails if they differ
[] Gitlab CI/CD pipeline to build a docker image
[] Published 0.0.1 version of the docker image