Metrics do not have instance_id or container_id in case when deployed using docker or kubernetes

[architgarg95]

What version of OpenTelemetry are you using?

"@opentelemetry/api": "1.9.0",
"@opentelemetry/auto-instrumentations-node": "0.49.1",

What version of Node are you using?

Node 14.17.6 and Node 16.19.0 both versions tested

What did you do?

I have multiple instances of a service running on 2 AWS EC2 (2 instances). Due to this the counter values maintained for http_server_duration_milliseconds_count have some difference that leads to Grafana Alloy (based on opentelemetry collector) showing a zig zag pattern (due to time diff between the 2 instances start). In java and python for multiple instances getting an instance__id associated with metrics but in case of Nodejs that was not the case.
Thought might be due to docker container or k8 cluster, tried the same for this application but still not getting any containerid or instance_id to distinguish between different instances metrics

What did you expect to see?

expected to see behaviour similar to Java and Python Opentelemetry wherein getting the instance_id attribute in the metrics

What did you see instead?

no instance_id attribute in metrics for nodejs

Additional context

This really affects the graph in grafana alloy resulting in rate graphs having a much higher value than showing zero, which really put us off

[architgarg95]

using node auto instrumentation, version specified above as well. Any help would be appreciated

[pichlermarc]

It look like there's a feature enabled on the Python and Java Prometheus exporters that makes resource attributes appear on the metric. This is an optional feature that we have not implemented yet, so I'm changing this to a feature-request.

Specification: https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/metrics/sdk_exporters/prometheus.md#configuration