Matthew James Briggs

Kubernetes provides built-in mechanisms for ownership referencing and garbage collection. Should we be using this for tests and/or resources when their controller is deleted? Should tests 'own' resources?

https://github.com/bottlerocket-os/bottlerocket-test-system/pull/139/files#r739505986 - Do secrets of different types already work fine? - If not, do we need to support secrets of other types? - If No/Yes: make the necessary changes.

As described in #428, having errant nodes in the cluster would produce incorrect conformance testing results. Additionally, if a conformance testing process failed without proper cleanup, it is possible for...

Tangentially related to #353, I had an error that I think resulted because my resource name had a dot in it. But more importantly, it is essentially impossible (?) to...

We have discovered that GitHub archive hash sums can change. We should probably move all of the sources we pull into the lookaside cache. When we do, we should switch...

See https://github.com/bottlerocket-os/bottlerocket/blob/develop/packages/aws-iam-authenticator/aws-iam-authenticator.spec#L20. _Originally posted by @jpculp in https://github.com/bottlerocket-os/bottlerocket-test-system/pull/283#r799786300_

Let's use this as a tracking issue for things that need to be implemented or fixed in the controller relating to handling errors. Background: The happy path of the controller...

When running sonobuoy tests, if nodes from some previous test case (an earlier version of Bottlerocket, for example) are running in the cluster in addition to, or in place of,...

Provide a cluster using AWS Cluster API. Run Sonobuoy using a Cluster API-provided EC2 cluster. - depends on #32 - depends on #38 `id: capa_cluster_provider`

It is possible for the resource controller to wait forever. At some point in the lifecycle, walk the dependency graph and raise an error if cycles are detected.