Skip to content

[exporter/prometheusremotewrite] 5xx errors incorrectly classified as error_permanent="true" #48431

@bogdan-st

Description

@bogdan-st

Component(s)

exporter/prometheusremotewrite

What happened?

Description

The PRW exporter reports all HTTP errors as error_permanent="true" on otelcol_exporter_send_failed_metric_points_total, including 5xx responses. Those are backend failures, not data errors, so you can't use error_permanent to tell the difference between the backend being down and a tenant sending out-of-order samples.

Steps to Reproduce

Set service.telemetry.metrics.level: detailed on a collector with a PRW exporter pointing at a backend that returns 5xx. Send any metrics through it. Check otelcol_exporter_send_failed_metric_points_total, error_permanent="true" regardless of status code.

Expected Result

error_permanent="false" for 5xx responses, error_permanent="true" for 4xx.

Actual Result

error_permanent="true" for everything.

Collector version

v0.152.0

Environment information

No response

OpenTelemetry Collector configuration

Log output

Additional context

No response

Tip

React with 👍 to help prioritize this issue. Please use comments to provide useful context, avoiding +1 or me too, to help us triage it. Learn more here.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions