We exported service A from cluster A and import it to cluster B for pod B to call service A.
When we do rollout restart for deployment A, pod B will encounter failed TCP connection issues to call service A, even though rollout restart shouldn't affect the availability of service A. The disruption will recover after a short while.
We noticed there are endpoints registering/de-registering operations from cloud-map-mcs-controller-manager during the restart, and we don't have Health check configured in cloud map, so some traffic may send to the old endpoints before it get de-registered, will it be the cause of connection disruption?
If yes, may I know if cloud-map-mcs-controller-for-k8s supports configuration of Health check? If no, can we request for this feature?
Thanks in advance and your answers will be much appreciated!
Hello,
We exported service A from cluster A and import it to cluster B for pod B to call service A. When we do
rollout restart
for deployment A, pod B will encounter failed TCP connection issues to call service A, even thoughrollout restart
shouldn't affect the availability of service A. The disruption will recover after a short while.We noticed there are endpoints registering/de-registering operations from cloud-map-mcs-controller-manager during the restart, and we don't have Health check configured in cloud map, so some traffic may send to the old endpoints before it get de-registered, will it be the cause of connection disruption?
If yes, may I know if cloud-map-mcs-controller-for-k8s supports configuration of Health check? If no, can we request for this feature?
Thanks in advance and your answers will be much appreciated!