Joy Nag

Addressed the feedbacks, PTAL 🙏

@daniel-nichter PTAL. I think I have addressed all your feedbacks. 🙏

We are wondering if something has changed between vector version ` 2023-01-27_backport_dd_metrics_interval_fix-distroless-libc` and `0.40.0` that could have caused this. As they were released ~1 yr apart, there must have been...

> Had you tried either of the following combinations: > - 2023-01-27_backport_dd_metrics_interval_fix-distroless-libc deployed as a daemonset > - 0.40.0 deployed as an aggregator We haven't tried it yet, I can...

@jszwedko Deployed `0.40.0` binary into the Aggregator deployment and the problem persisted. So it's not an problem arising from the switch from aggregator to daemonset. I then went on a...

Here's a snapshot of `0.26.0` introducing double tagging into the metric. You can see the metrics were correctly tagged and then suddenly a new series with double tag appears: ![Screen...

I will try `vector tap --outputs-of datadog_agent` and `tcpdump` on the DD agent. > One thing that might be useful is the Datadog Agent version. Could you provide that? We...

Tried `vector tap --outputs-of datadog_agent.metrics | grep --line-buffered "cpu\.usage\.total.*istio-proxy.*"` and found some metric data points that are having issues of multiple values for the service tag, and datadog-agent is indeed...

I tried putting in the following VRL check in one of our transformations: ``` if is_array(.tags.service) && .tags.container_name == "istio-proxy" { .tags.service = to_string!(.tags.service[-1]) } ``` And a test: ```...

Thanks for taking a look. I will go with `.tags = .tags` and `metric_tag_values: single` to unblock us for now while we work with Datadog Support to fix the root...