Make clusterName optional on EKS and GKE

[crobert-1]

Description:
Now that the resource detection processor supports detecting the k8s.cluster.name in GKE and EKS environments, the clusterName value can be optional in these environments.

A simplified resource detection processor has been added to more pipelines in this change as well, in every place the clusterName was manually being added.

Testing:
Tested in GKE and kOps. GKE sets cluster name properly without it specified, and overrides the detected k8s.cluster.name when clusterName is set. kOps fails without clusterName with the following message:

crobert$ ~/dev/splunk-otel-collector-chart $ helm upgrade ...
Error: UPGRADE FAILED: values don't meet the specifications of the schema(s) in the following chart(s):
splunk-otel-collector:
- (root): Must validate "then" as "if" was valid
- clusterName: String length must be greater than or equal to 1
- (root): Must validate all the schemas (allOf)

The - (root) lines are a bit confusing, but that's simply saying that the check failed, AFAICT.

Also, removed clusterName option from functional tests and changed expected k8s.cluster.name value in telemetry to be rotel-eks instead of the manually set value.

Documentation:
Updated README and relevant comments in code.

[crobert-1]

Note, this is a second attempt of #1056 to enable testing to work properly. I have not yet implemented changes requested in that review, other than README wording.

[crobert-1]

I believe I've addressed all comments on the original PR now, so I'm marking as ready for review.

[jvoravong]

One nit. Mostly LGTM. Approving. Please address other review comments before merging.

[jinja2]

@crobert-1 I think it is worth updating the eks functional_tests to test with the resourcedetector added cluster-name. The test currently sets the cluster name in values which if removed, should default to the resourcedetector provided attr. This would need a change to the testdata, since the name of the eks cluster being tested on is rotel-eks and not the provided dev-operator.

[crobert-1]

I've updated functional tests to account for the EKS optional cluster name. These tests are currently being skipped as the functionality is not complete, but the logic is in place for when they're running. I've moved the kind cluster test data to a sub-directory, along with the eks cluster data to its own sub-directory. The EKS data is not yet being used.

[jinja2]

hmm, weird that the EKS tests are being skipped after the rebase since the source isn't a fork nvm, the eks tests were skipped this time due to kind tests failing.

[crobert-1]

Regarding the failing upgrade test: It's failing because the cluster name is no longer passed in to the install command. Since the test is running a previous version of the chart without this PR's changes, the clusterName is still required, even in EKS environments. I think it's okay to merge this change as once this PR is no longer the latest commit in main, it should pass again.

[atoulme]

Please go ahead and merge.