Sure, I can provide some guidance based on the information available in the extracts. It seems like you're experiencing issues with the Liveness/Readiness probes in your EKS deployment. This issue was discussed in a [previous support conversation](https://airbytehq.slack.com/archives/C021JANJ6TY/p1676909154510159) on Slack. The user in that conversation was also experiencing frequent restarts of the server component due to timeouts in the Liveness/Readiness probe. One of the suggestions was to disable the liveness probes and only keep the readiness ones. Another user suggested that the health check endpoint, which reads from the database as part of the health check, might be causing the issue. They proposed changing the

SELECT id FROM workspace LIMIT 1

to

SELECT 1

in

io.airbyte.config.persistence.ConfigRepository

. Here's the proposed change:

java database.query(ctx -> ctx.fetch("SELECT 1")).stream().count();

This change is meant to reduce the load on the database by simplifying the health check query. Another suggestion was to ensure that readiness and liveness probes do not depend on downstream systems like Postgres. This could be done as part of the startup probe to prevent misconfigured pods from being considered OK. Please note that these are suggestions based on previous discussions and may not directly solve your issue. If the problem persists, I recommend reaching out to the Airbyte community on [Slack](https://slack.airbyte.com/) for further assistance.