Insights

Field reports from the trenches

Every post here describes real work — including the mistakes and fixes your team will benefit from. Client names are never mentioned.

Apr 26, 2026· 6 min read

SLO burn rates without a vendor — in 400 lines of JS

How we built a multi-window error-budget engine that reads Prometheus, per the Google SRE workbook — without Datadog or Honeycomb. With the actual code.

SLOburn ratePromQLGoogle SRE

Apr 26, 2026· 8 min read

OpenTelemetry, Jaeger, Prometheus, Grafana, SLOs — in one sprint

A play-by-play of going from zero observability to traces in Jaeger, 52 metrics in Prometheus, 5 Grafana dashboards, and an SLO catalogue with multi-window burn-rate alerts — without taking the platform down.

observabilityOpenTelemetrySLOGrafana