When changing cluster node type Qovery shows a warning:
Changing these parameters might cause a downtime on your service.
Could you link to documentation and elaborate more on when can you expect down time? It seems like new nodes could spin up before destroying old ones letting them drain over time until all pods get moved to the new nodes.
Hi @prki,
we don’t have documentation on it, I’ll try to summarize it here and then report it on a doc.
Node updates are managed one by one: a node with the new configuration is created, and an “old node” is drained (i.e. shut down)
The impact that the node drain operation has on your service mostly depends on your application setup and current load, I’ll try to summarize it here:
– resource issues: if your app is already running with X instances, where X = max allowed number of running instances for the app, removing one instance might increase the load on the remaining instances (and maybe service disruption for some clients of that service)
– Availability issues: if your app is running with only 1 instance (no matter the min/max instance configured), Kube can still remove only one instance running to drain the node. In this case, we suggest having at least min = 2
I hope this helps
This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.