Fix metric mapping of counter metrics in the Openmetrics V2 version of the check

[steveny91]

What does this PR do?

Counters weren't getting collected because Openmetrics v2 expects counter metrics to be specified without the _total. This PR removes _total from all the counter metrics in the openmetrics version of the check. This check never had a openmetrics v1 implementation. So I figured I can edit the main metric dictionary.

Also fixes one metric that was being mapped to the wrong datadog name envoy_cluster_upstream_cx_tx_bytes_buffered

[github-actions]

The validations job has failed; please review the Files changed tab for possible suggestions to resolve.

[github-actions]

The validations job has failed; please review the Files changed tab for possible suggestions to resolve.

[github-actions]

The validations job has failed; please review the Files changed tab for possible suggestions to resolve.

[codecov]

Codecov Report

Merging #13573 (49ecebd) into master (4c20ed0) will not change coverage.
The diff coverage is n/a.

Flag	Coverage Δ
envoy	94.23% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

[alopezz]

Do we have an answer to extra_metrics possibly not being honored (as mentioned in #12855 (comment))?

[steveny91]

Do we have an answer to extra_metrics possibly not being honored (as mentioned in #12855 (comment))?

I tested it and I didn't have an issue adding an extra metric. For example:

# TYPE envoy_cluster_upstream_cx_http1_total counter
envoy_cluster_upstream_cx_http1_total{envoy_cluster_name="dummy_dynamic_cluster"} 0

with:

    extra_metrics:
      - envoy_cluster_upstream_cx_http1:
          name: test

I get:

$datadog-agent check envoy --table -r | grep test                            ok  7s  10:17:39 AM 
  envoy.test.count                                                         count  1672928266  0                     endpoint:http://localhost:8001/stats/prometheus, envoy_cluster:dummy_dynamic_cluster                                        

I can't reproduce the issue.