Is it possible to run the copy analyzer against multiple namespaces, with a wildcard character (for instance, “my-namespace-prefix-*”) instead of against a hard-coded namespace? Or, alternatively just tell it to run in all namespaces within the cluster?
As part of our application, users can run “deployments” that create namespaces and pods within the cluster on their own. We know the prefix of the namespaces, but the final “slug” of the namespace is entirely defined by the user. We know what gets deployed into each of these namespaces, so we have a proper selector we can use, but no way of knowing if we can run the copy collector without hard-coding the namespace.
That’s not currently in the product, it sounds like a very interesting use case that we’ve not come across previously.
Perhaps the thing to do is provide a more detailed description of what the required outcome is in a feature request over at the Troubleshoot project repo where the team can review what it would need to look like to implement.
I followed what was mentioned in this other topic, and this worked for us:
Basically, we just don’t set the namespace: for the copy collector and used a selector that was shared across the deployments across multiple namespaces and we were able to collect the logs across namespaces.