vortegatorres commented on issue #13282: URL: https://github.com/apache/apisix/issues/13282#issuecomment-4439725859
Hi! We are hitting the same issue in production (APISIX 3.16.0 standalone, `lua-resty-healthcheck-api7 3.2.1-0`, upstreams using `service_name + discovery_type: "kubernetes"`): long-lived APISIX pods keep health-checking Pod IPs that are no longer present in either the standalone admin `/configs` output or the `/v1/discovery/kubernetes/dump` output, until the pod is restarted. We previously reported the same underlying behaviour from the ingress-controller side in apache/apisix-ingress-controller#2708, where it surfaced as stale IPs lingering in `upstream_addr`; with active health checks enabled it now also produces `lua-resty-events` queue-overflow floods on the affected pods. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
