Mikołaj Świątek

To be clear, we can merge your fix without figuring this out, but the discrepancy suggests we're doing something differently than Prometheus, which is a cause for concern for me.

Allright, that makes sense. Since target allocator uses a different role, let's fix it in some deterministic way, and once there's an upstream solution, we can switch to that. Thank...

Sounds like it should be updated, indeed. Your proposed fix is misleading, though: ```go // Allocation strategies supporting high availability: // - consistent-hashing: Supported (distributes targets via hashing) // -...

Is there an idiomatic mechanism to represent this kind of relationship in Kubernetes? I'm not sure labels are the right one. As Pavol noted, we do have plans to use...

This is not an unreasonable request, but it creates a lot of problems if we follow the logic to its conclusion. Why should the annotations be propagated only to the...

I'm much more open to the idea of making the behavior consistent between the OpenTelemetryCollector and TargetAllocator CRs, even if it's somewhat arbitrary. If MutatingAdmissionPolicy is not something that's available...

Looks like you have some test failures @mikel-jason, can you have a look?

I don't see any reason not to add this, we already have a few Service-specific attributes in the common part of the CRDs. If anyone wants to contribute this, we'll...

Would the [TargetAllocator CRD](https://github.com/open-telemetry/opentelemetry-operator/tree/main/cmd/otel-allocator#targetallocator-crd) not solve this problem? You can set volumes and volume mounts on it independently of the otel collector.

The PR looks great, thanks for the contribution! There's one big problem with it, and to be clear, this is really Kubernetes' fault. As it turns out, the OpenAPI spec...