feat: expose prometheus nodeselector and tolerations in monitoringstack CRD #293
Conversation
|
If this PR changes the Prometheus |
@simonpasquier, I like the idea to have a global setting for all child resources, however should we also allow each child resource to have different setting to the global one? |
@vimalk78 shouldn't that test fail in that case 🤔 |
it would, but the new parameters are optional, so unless explicitly set in the object the managedFields wont change. |
|
@sthaha my gut feeling is that in most cases, you want the monitoring pods to run on the same set of nodes (e.g. infra) so having a top-level field is probably what makes the most sense. We can argue later if there's a need for individual settings for each component but I'd rather make the mainstream situation easy to configure. |
|
@simonpasquier +1 to having it global for now
+1 to YAGNI :) |
| NodeSelector: config.NodeSelector, | ||
| Tolerations: config.Tolerations, |
There was a problem hiding this comment.
@bng0y , Like @simonpasquier suggested, lets make this global and apply the same to all resources that ObO creates.
|
@bng0y , It would be great to have this feature added to ObO. Do you need help implementing this? |
Allows to isolate the Prometheus pods when necessary.
Context: https://issues.redhat.com/browse/SDE-2823