[metricstarttimeprocessor] Identify resource by hash of all attributes

[dashpole]

Component(s)

metricstarttime processor

Describe the issue you're reporting

Currently, the implementation of true reset points in the prometheusreceiver only uses service.name and service.instance.id to identify a resource:

opentelemetry-collector-contrib/receiver/prometheusreceiver/internal/metrics_adjuster.go

Line 268 in 6ea0ea7

_, found := rm.Resource().Attributes().Get(semconv.AttributeServiceName)

This logic is being copied to the metricstarttime processor in #37855. However, this logic is prometheus-specific, and should use a hash of all resource attributes instead of just those two.

Possible solutions

This identity library may be helpful here: https://github.com/open-telemetry/opentelemetry-collector-contrib/tree/6ea0ea7e0423dd5271cc4ef1b6bde21d6f598412/internal/exp/metrics/identity.

Alternatively, we could use hash/fnv, similar to https://github.com/GoogleCloudPlatform/opentelemetry-operations-go/blob/19c4db6ea12211308fbd2cba12cc8665a5b7c890/exporter/collector/internal/datapointstorage/datapointcache.go#L250. This is similar to how the prometheus server constructs hashes of labels.

[github-actions]

Pinging code owners for processor/metricstarttime: @dashpole @ridwanmsharif. See Adding Labels via Comments if you do not have permissions to add labels yourself. For example, comment '/label priority:p2 -needs-triaged' to set the priority and remove the needs-triaged label.