Error: 14 UNAVAILABLE: read ECONNRESET

[vasu-rangpariya-yudiz]

Problem description

When I am trying to call a grpc request from a grpc client it works well for nearly 30-40 mins but after that time it automatically throws an error about the backend being unavailable although grpc server backend is running fine. after 2-3 failed requests It will automatically reconnect the grpc server and the same case happens again and again

Reproduction steps

Create a grpc server, a client with a docker image of 20-alpine and add corn that automatically calls one rest API and that rest API calls the grpc server after 30 minutes it gives a connection error

Environment

• Linux, 5.10.199-190.747.amzn2.x86_64
• Node v20 with Kubernetes with AWS ALB
• @grpc/grpc-js, v1.9.13 [e.g. gRPC@1.12.0]

[murgatroid99]

It would help to get client trace logs of one of these errors. Trace logging can be enabled by setting the environment variables GRPC_TRACE=all and GRPC_VERBOSITY=DEBUG. If you want to redirect the logging output you can use grpc.setLogger.

Specifically, it would help to see continuous logs starting with a request that succeeds, and then a request that fails with this error, and then another request that succeeds.