Nic Cope

> This PR does two things @sttts +1 for deprecating the `APIUpdatingApplicator`. That makes sense to me. With regard to the `APIPatchingApplicator` changes, my understanding (as written up in https://github.com/crossplane/crossplane-runtime/issues/483#issuecomment-1636432560)...

> So the edge case we'd be susceptible to would be hitting the same error twice in a row. In that case nothing would trigger a reconcile to be queued...

Ping. Any thoughts on this?

Would a new `PushSecrets` (plural) type would be a good approach here? The general idea is that `PushSecret` remains as-is - it pushes a single, named secret. `PushSecrets` is a...

Just chiming in - I run the zipkin collector in a Kubernetes cluster and I'm using the `/health` endpoint for liveness and readiness probes without issue. It would be nice...

I think this is a duplicate of https://github.com/crossplane/crossplane/issues/1699, though I'm tempted to keep this newer issue as it has a little more detail. This is also loosely related to https://github.com/crossplane/crossplane/issues/2255....

Is there any chance we could get away with just lumping this in with the existing `deletionPolicy: Orphan`? i.e. Always leave and orphan the connection secret if the managed resource...

Definitely worth considering. I don't feel strongly here - just looking for places to keep our API a little simpler. I like the idea of being able to handle this...

> Does this mean it won't be possible to have a fluent migration from v1 to a possible future v2 Composition? I'm not familiar with the term "fluent migration", so...

> However, package revision types are somewhat of an "active" resource with a controller in the background that reconciles it and has an effect on the cluster. A CompositionRevision on...