Skip to content

RUNBOOK-tracecore.md: process_resident_memory_bytes is not emitted; replace with otelcol_process_memory_rss #294

Description

@trilamsr

Summary

docs/RUNBOOK-tracecore.md line 51 + 65 (and any other use) cites
process_resident_memory_bytes{job="tracecore"} as the operator query
for collector memory. That metric is the prometheus-client-golang
default process collector
convention; upstream OTel service.telemetry
does not register it. The actually-emitted metric is
otelcol_process_memory_rss (per
docs/migration/v0.1-to-v0.2.md line 98 — already documented).

Impact

Operators following the runbook during an OOM incident paste a query
that returns no data. The dashboard panel built from the runbook query
shows "No data" right when the incident is being debugged.

Fix

Replace the metric name in docs/RUNBOOK-tracecore.md "Collector pod
OOMKilled" section:

process_resident_memory_bytes{job="tracecore"} / 1024 / 1024

otelcol_process_memory_rss{job="tracecore"} / 1024 / 1024

And update the "chart" reference in the remediation block (line 65)
accordingly.

Verification

grep otelcol_process_memory_rss internal/integration/ocb_scrape_test.go
should be extended to assert the metric ships on a live OCB scrape.

References

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions