← Back to guidesGuide

Reduce Prometheus cardinality spikes with OpenTelemetry limits before the TSDB burns

A practical workflow to detect explosive series, understand OpenTelemetry overflow, and trim labels without breaking useful alerts.

Learn how to find metrics that create runaway cardinality, separate SDK protection from real Prometheus cost, and apply safe mitigations in instrumentation, the Collector, and scrape config.

Created: August 28, 2026

Published: August 28, 2026

Estimated time45 min

LevelAdvanced

Before you startRead access to Prometheus or a PromQL-compatible backend

PlatformsLinux

WhatsApp X LinkedIn

Linux

Diagnose from a shell with the Prometheus HTTP API, optional promtool, and local Collector configuration.

curljqoptional promtoolaccess to the collector config

Top series by metric and job

curl -G http://prometheus:9090/api/v1/query --data-urlencode 'query=topk(20, count by (__name__, job)({__name__!=""}))' | jq '.data.result[] | {metric: .metric, value: .value[1]}'

Search for OpenTelemetry overflow

curl -G http://prometheus:9090/api/v1/query --data-urlencode 'query=count by (__name__, job)({otel_metric_overflow="true"})' | jq '.data.result'

Also available inReducir picos de cardinalidad en Prometheus con límites de OpenTelemetry antes de quemar la TSDB

Content locked

This guide requires both steps before full content is available.

○Click “Like” on this guide.
○Share on WhatsApp, X, LinkedIn, or copy the link.

Access is automatically unlocked as soon as both steps are completed.