Posts tagged prometheus

A Hands-On Example of Decoupling App Instrumentation with OpenTelemetry

Blog Post Publish Date: 2026/11/18

This post builds a small Python service instrumented with OpenTelemetry, one piece at a time. It starts with the API only, turns on the SDK without changing the code, routes the metrics through the OpenTelemetry Collector to Prometheus, and then switches to Datadog by changing only the Collector configuration.

Read more ...


Stop Extracting Metrics from OpenSearch Logs

Blog Post Publish Date: 2026/11/03

This blog post explains why counting log lines in OpenSearch is an expensive way to produce a metric. The same number captured as a Prometheus counter is stored once and read almost for free. When you cannot change the application, Fluent Bit can extract that metric while the logs are ingested, so OpenSearch never has to rebuild it at query time.

Read more ...


OpenTelemetry: A Framework for Decoupling Your Observability

Blog Post Publish Date: 2026/10/29

OpenTelemetry is not a database, and it doesn't compete with Fluent Bit, Logstash, or Jaeger. It defines how metrics, traces, and logs are structured and transported, so your telemetry isn't tied to any vendor. The tools you already know can be the backends that receive it. This post explains that idea, the one component OpenTelemetry runs on its own, and where it fits.

Read more ...