Staging instance, all changes can be removed at any time

Skip to content

k8s: Speedup the deployments

Vincent Sellier requested to merge k8s-deployments-speedup into master

It's a workaround for the "failed to reserve container name" issue during big deployments. Most of the pods fails to start and takes a lot of time to recover.

The default value is 2mn which is too small when a big deployment is in progress.

A lot of information about this issue is available at: https://github.com/containerd/containerd/issues/4604

This MR is for the record. It's already applied on all the clusters.

Production recovered in 15mn while it failed to deployed the last deployment since a couple of hours.

Edited by Vincent Sellier

Merge request reports