← sources

Kubernetes probes keep restarting services that just need more time

source: reddit (reddit:devops)published 2026-10-20 14:32ingested 2026-10-21 02:37

retained text

Hey folks, I've noticed readiness and liveness probes often end up with pretty similar settings, even tho they do different jobs. With tight timeouts and failure thresholds, a service that's slow to start or recovering from a dependency issue can get pulled out of rotation or repeatedly restarted.

I've seen probe settings get copied between services too, even when their startup times are completely different.

How do you guys set probe thresholds for services with unpredictable startup times or dependencies that occasionally respond slowly?

document fetched 2026-10-21 00:24 · canonical https://reddit.com/r/devops/comments/1xatqcs/kubernetes_probes_keep_restarting_services_that