Skip to main content
Target Allocator distributes Prometheus scrape targets across OpenTelemetry Collector replicas. It keeps ServiceMonitor and PodMonitor scraping balanced as collector replicas and monitored targets change.

​
Talks To

​
How It Fits

Target Allocator watches Kubernetes resources used for Prometheus-style discovery, including ServiceMonitor and PodMonitor objects. It combines those resources with any explicit scrape configuration rendered into the chart values. Collectors call the target allocator to get their assigned scrape targets. This keeps collectors from all scraping the same targets and lets scrape ownership move when collector replicas are added, removed, or restarted. Target Allocator does not forward telemetry to the hub. It controls scrape distribution; the OpenTelemetry Collector performs the scrape and Metoro Exporter forwards the resulting metrics.

​
Operational Notes

The target allocator is stateless and runs as a Deployment. Its configuration comes from the chart-rendered ConfigMap and Kubernetes API watches. If it is unavailable, collectors may be unable to refresh scrape assignments. That typically appears as missing or stale scraped metrics rather than a failure in node-agent traces, logs, or profiles.