Replace the monitoring loop, not just the dashboard

Replace the monitoring loop, not just the dashboard

Replacing a monitoring dashboard isn’t the same as replacing its production feedback loop.

In “How to build an open-source Datadog,” I outline a setup with OpenTelemetry as the instrumentation contract, SigNoz as the central backend, and an external uptime monitor. Keeping instrumentation separate from the backend makes that backend replaceable; checking uptime from outside the application adds a different view of whether it’s available.

That’s the useful test for an observability stack: not just what it displays, but how it helps you notice and investigate a production problem.

https://grnmn.com/blog/how-to-build-an-open-source-datadog/