We got quite some instabilities on our Airbyte instances (dev/prod), we found why pods for minio/postgres can’t get enough resources (CPU/RAM) on the kube cluster, problem we can’t set the resources on the minio pod and the PostgreSQL pod currently (due to the way their helm chart is structured). We see as alternatives: 1 - We submit a patch upstream, wait for it to get merged, upgrade, done 2 - We move to GCS/cloudsql, a bit of pain, but we don’t have to care about backups/reliability anymore Are we on the good track ? Do you see others alternatives ? We are probably to go for number 2 unless you have another tip for us 🙂 / @kapa.ai

New message text here

solution 1 : JOB_MAIN_CONTAINER_MEMORY_REQUEST is only for jobs sync, not for minio/postgres so this doesn’t solve our issue. solution 2 => doesn’t solve either 3 and 4 are our solutions suggested