This message was deleted.

Are you running MongoDB in a replica set with multiple members? That's the recommended approach for prod to avoid server selection errors. As for pod health checks, there are a couple options. 1. Run front-end and back-end as separate pods. Same image, but the commands would be different:

yarn workspace front-end start

and

yarn workspace back-end start

. If you do that, if the server dies, the whole pod should stop 2. The back-end has a

/healthcheck

endpoint. For the front-end, you can use the

/api/init

endpoint.

Currently there is only a single pod for MongoDB and a single pod for the main application. Thanks for the suggestions! Will try splitting up our deployments into separate front-end and back-end pods with individual health checks.