We are trying to install a self-hosted instance of AppSmith Enterprise Edition on our AWS EKS cluster using the helm chart v3.6.3 (which uses the "appsmith/appsmith-ee:latest" container image) and the app appears to crash before even reaching the initial first user sign up screen.
The interesting difference with our setup is that we would like to host our AppSmith with managed infrastructure; AWS RDS for Postgres, AWS Elasticache for Redis, and AWS DocumentDB for MongoDB, instead of through the built-in dependency charts.
Digging through the logs, I suspect that there is an issue with the "temporal" service. It doesn't seem to start up properly and starts churning out errors like the following:
temporal stdout | {"level":"error","ts":"2025-07-22T20:06:58.961Z","msg":"Unable to call matching.PollWorkflowTaskQueue.","service":"frontend","wf-task-queue-name":"appsmith-0:62d438b1-9557-4d
aa-83bb-2b59dc09281c","timeout":"1m9.984765798s","error":"Not enough hosts to serve the request","logging-call-at":"workflow_handler.go:918","stacktrace":"go.temporal.io/server/common/log.(*za
pLogger).Error..."}
A copy of our helm chart values:
redis:
enabled: false
mongodb:
enabled: false
postgresql:
enabled: false
customCAcert:
"global-bundle.pem": "***" # used for documentdb tls connection reasons
applicationConfig:
APPSMITH_DB_URL: "mongodb://user:pass@mongodb_host:27017/appsmith?ssl=true&replicaSet=rs0&readPreference=secondaryPreferred&retryWrites=false&tlsCAFile=/appsmith-stacks/ca-certs/global-bundle.pem.crt"
APPSMITH_REDIS_URL: "redis://appsmith-redis.****.0001.usw2.cache.amazonaws.com:6379"
APPSMITH_KEYCLOAK_DB_DRIVER: "postgresql"
APPSMITH_KEYCLOAK_DB_USERNAME: "databaseuser"
APPSMITH_KEYCLOAK_DB_PASSWORD: "******"
APPSMITH_KEYCLOAK_DB_URL: "postgresql://user:pass@postgres_host/appsmith?sslmode=require"
SQL_TLS_ENABLED = true
SQL_TLS_DISABLE_HOST_VERIFICATION = true
SQL_TLS = true