I have been getting server errors when trying to deploy my app in Langgraph Cloud. These errors occur whether I use the web-based deployment or the Control Plane API. The logs look like this:
18/09/2025, 13:57:27 [WARNING] error connecting in 'pool-1': [Errno -2] Name or service not known
18/09/2025, 13:57:26 [WARNING] error connecting in 'pool-1': [Errno -2] Name or service not known
18/09/2025, 13:57:25 [WARNING] error connecting in 'pool-1': [Errno -2] Name or service not known
18/09/2025, 13:57:24 [WARNING] error connecting in 'pool-1': [Errno -2] Name or service not known
18/09/2025, 13:57:24 [INFO] Starting Postgres runtime with langgraph-api=0.4.20
18/09/2025, 13:57:24 [INFO] Waiting for application startup.
18/09/2025, 13:57:24 [INFO] Started server process [1]
18/09/2025, 13:57:24 [INFO] Loaded custom auth middleware: CustomAuthBackend(fn=<function validate_auth0_token at 0x7fe5611e2e80>, ls_auth=<langgraph_api.auth.langsmith.backend.LangsmithAuthBackend object at 0x7fe5611be6d0>, param_names={'authorization', 'headers'})
18/09/2025, 13:57:24 [INFO] Loaded auth instance from path /deps/app/src/security/auth.py:auth: <langgraph_sdk.auth.Auth object at 0x7fe5612e04a0>
18/09/2025, 13:57:24 [INFO] Using custom authentication
18/09/2025, 13:57:24 [INFO] Using auth of type=custom
18/09/2025, 13:57:24 [INFO] Using langgraph_runtime_postgres
18/09/2025, 13:57:23 [WARNING] error connecting in 'pool-1': [Errno -2] Name or service not known
18/09/2025, 13:57:23 [INFO] Starting Postgres runtime with langgraph-api=0.4.20
18/09/2025, 13:57:23 [INFO] Loaded custom auth middleware: CustomAuthBackend(fn=<function validate_auth0_token at 0x7e3d43064a40>, ls_auth=<langgraph_api.auth.langsmith.backend.LangsmithAuthBackend object at 0x7e3d4306b4d0>, param_names={'headers', 'authorization'})
18/09/2025, 13:57:23 [INFO] Loaded auth instance from path /deps/app/src/security/auth.py:auth: <langgraph_sdk.auth.Auth object at 0x7e3d43a01940>
18/09/2025, 13:57:22 [INFO] Using custom authentication
18/09/2025, 13:57:22 [INFO] Using auth of type=custom
This error occurs many times as the deployment retries, and is eventually replaced by:
18/09/2025, 14:05:40 [WARNING] error connecting in 'pool-1': connection failed: connection to server at "172.26.74.14", port 5432 failed: FATAL: password authentication failed for user "postgres"
18/09/2025, 14:05:40 [WARNING] error connecting in 'pool-1': connection failed: connection to server at "172.26.74.14", port 5432 failed: FATAL: password authentication failed for user "postgres"
18/09/2025, 14:05:39 [WARNING] error connecting in 'pool-1': connection failed: connection to server at "172.26.74.14", port 5432 failed: FATAL: password authentication failed for user "postgres"
18/09/2025, 14:05:39 [INFO] Starting Postgres runtime with langgraph-api=0.4.20
18/09/2025, 14:05:39 [INFO] Waiting for application startup.
18/09/2025, 14:05:39 [INFO] Started server process [1]
Note that I am not attempting (consciously) to use my own Postgres instance. This is a fully cloud-based deployment. I set a number of secrets in the environment, but none of them are related to Postgres.
Is this caused by something wrong in my deployment? It doesn’t feel like this error is something that I have any control over, but perhaps I’ve misunderstood something.