OpenTelemetry at Home: Traces, Metrics, Logs

Thu, 09 Jul 2026 06:00:00 +0000

For years my homelab observability was three unrelated tools that happened to share a Grafana dashboard. Prometheus scraped metrics endpoints. Loki tailed logs. And when a request was slow, I had no idea why, because nothing told me which service in the chain was actually the bottleneck — I’d grep logs across four containers by timestamp and hope the clocks agreed. That last part is what traces are for, and it’s the piece most homelabs never bother with, because setting up distributed tracing used to mean picking a vendor-specific agent and hoping it played nicely with everything else.

Opentelemetry - vo.rs

OpenTelemetry at Home: Traces, Metrics, Logs