How-to guideobservability10–12 minConceptual
Defining SLOs that Actually Drive Alerts
Use SLOs to decide when to alert and when to wait.
SRELast updated 2025-11-26
sloserror-budget
Share:
What makes a good SLO?
- Expressed in terms your users feel: availability, latency, correctness.
- Bounded over a rolling window, e.g. 30 days.
- Backed by clear measurement from logs, traces, or metrics.
Example availability SLO
A classic starter SLO is: > 99.9% of successful requests to `/api/*` over 30 days You can derive this from HTTP status codes and latency metrics in your existing observability stack, then connect alerts into TG2G via the Event API or one of our alert integrations.