[backend] metadata-grpc-service connection error: connection reset by peer

[fantasy111]

Environment

• How did you deploy Kubeflow Pipelines (KFP)?
  manifests
• KFP version:
  2.3.0
• KFP SDK version:
  2.11.0

Steps to reproduce

I write a simple python code as below:
@dsl.component()
def sleep_task(metrics: Output[Metrics]):
import time
time.sleep(1200)
metrics.log_metric("total", 1)

@dsl.pipeline(name='A2')
def my_pipeline():
sleep_op = sleep_task()

The kfp always output : F0221 07:45:04.241460 18 main.go:49] failed to execute component: unable to produce MLMD artifact for output "metrics": rpc error: code = Unavailable desc = error reading from server: read tcp 10.197.0.233:37640->10.198.43.0:8080: read: connection reset by peer
time="2025-02-21T07:45:04.328Z" level=info msg="sub-process exited" argo=true error=""
Error: exit status 1

Are there any parameters related to idletimeout that I need to set ？？？

Expected result

exit with code 0

Materials and Reference

---

Impacted by this bug? Give it a 👍.