0

We have a K8s Cluster (3 Master - 2 Worker) - v1.17

There are 2 Microservice in this cluster, a Microservices A call to Common. Sometimes, I face the problem is: A call to Common has timeout after 60s - Although this request is processed very quickly in the Common and success ( < 10ms).

getErrorInfoFallback : feign.RetryableException: Read timed out executing GET http://common-service.dev.svc.cluster.local:8002/errormapping/v1.0?errorCode=abcxyz

I use FeignClient to call other Microservice with url like http://common-service.dev.svc.cluster.local:8002

Here is timeline:
- 16:37:42.362 A send request 
- 16:37:42.368 Common logging the request 
- 16:37:42.378 Common logging respone return 
- 16:38:42.424 A: timeout exeption

Could anyone help me?

user3611168
  • 335
  • 1
  • 6
  • 27
  • How many pods are under service Common? Maybe port is being used or load on the pod under the service is high? – Crou May 12 '20 at 12:21
  • There are 4 pods of Common, 2 pods of A Service. I only testing with 20 concurent threads by jmetter. I was update the timeline, It's seem the request send to Common successful but response is not OK. – user3611168 May 12 '20 at 14:33
  • Have you checked metrics inside the `pod`? Also how exactly are you testing it and which CNI are you using? – Crou May 21 '20 at 15:00
  • It's me. CNI is flannel, I have take tcp dump and found a lot of TCP packet retranmission, and then TCP connection has been closed after some retry. – user3611168 Sep 17 '20 at 08:29

0 Answers0