sxcooler edited a comment on issue #1236: GET /v4/default/registry/instances ... Connection was closed URL: https://github.com/apache/servicecomb-java-chassis/issues/1236#issuecomment-510469818 I don't really know the reason of this issue, but here are some suggestions: First, set timeout like above. then, make sure you've set publishAddress as k8s ingress address, as default servicecomb will use pods address, like "foo-bar-99788d85d-8nqx5", this will cause some error when your pods scaling. Now I've got 2-3 "GET /v4/default/registry/instances ... Connection was closed" per day, I think it's acceptable. ``` servicecomb: ... service: publishAddress: foo-bar request: timeout: 30000 ``` And in your gateway's config (I'm using zuul): ``` zuul: ... host: maxTotalConnections: 1000 maxPerRouteConnections: 500 ribbon: eureka: enabled: false eager-load: ConnectTimeout: 5000 ReadTimeout: 15000 SocketTimeout: 15000 MaxTotalHttpConnections: 1000 MaxConnectionsPerHost: 500 hystrix.command.default.execution.isolation.thread.timeoutInMilliseconds: 60000 ``` Closing this issue for now.
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services
