We had been having pages lately due to zotero congestion and eventual restarts, creating a proper readiness probe should help with those pages and will instruct kubernetes to restart the pod if the service is unresponsive.
Current proposals to achieve this:
- Create a swagger spec definition for zotero and check the service using service-checker.
- Do a curl to local service passing a well known url for citations and check output.
- Create a proper health endpoint in the service.
The easiest low hanging fruit seems to be the curl so we will start there