OpsCanary
kubernetesobservabilityPractitioner

OpenTelemetry Migration: Scaling Your Metrics Platform in Kubernetes

5 min read CNCF BlogSep 17, 2026Reviewed for accuracy
Share
PractitionerHands-on experience recommended

In today's cloud-native landscape, standardizing on OpenTelemetry is crucial for effective observability. As more applications emit OTel data, the need to migrate existing metrics platforms becomes apparent. This migration not only streamlines data collection but also enhances compatibility with modern tools, ultimately improving your monitoring capabilities.

The migration process is methodical. First, replace the gostatsd sidecar with your OTel Collector distribution, ensuring that applications continue to send StatsD data over UDP as they did before. For ingestion, utilize the contrib loadbalancingexporter, which can hash by streamID, optimizing data handling. The aggregation process is impressive; you can handle approximately 4.8 billion data points per minute, reducing that to around 220 million—an astounding 96% reduction. Finally, transition to a stateless Collector distribution, known as metrics-gateway, which is built on upstream exporters, allowing for efficient data forwarding.

In production, be aware of the scale of your data. The migration to OpenTelemetry is not just about adopting a new tool; it’s about understanding how to optimize data flow and aggregation. This process can significantly enhance your metrics platform, but it requires careful planning and execution to ensure that you don’t lose critical insights during the transition.

Key takeaways

  • Replace the gostatsd sidecar with an OTel Collector distribution for seamless migration.
  • Utilize the contrib loadbalancingexporter to hash by streamID for better data ingestion.
  • Achieve a 96% reduction in data points by optimizing your aggregation process.

Why it matters

Migrating to OpenTelemetry can significantly enhance your observability stack, allowing for better insights and performance monitoring. This transition not only reduces data volume but also aligns your metrics collection with industry standards.

When NOT to use this

The official docs don't call out specific anti-patterns here. Use your judgment based on your scale and requirements.

Want the complete reference?

Read official docs

Test what you just learned

Quiz questions written from this article

Take the quiz →
Linux Foundation🔥 SEPTEMBER PROMOSponsor

35% off certifications and e-learning with code SEPT26BTS35, or 40% off bundles and instructor-led training with SEPT26BTS40. New this month: the MCPA (Model Context Protocol Associate) certification.

Claim the discount →

Get the daily digest

One email. 5 articles. Every morning.

No spam. Unsubscribe anytime.