Linux Labs

SRE practices: SLOs, incidents, redundancy / Stage 1 of 8

SLI: the share of good requests

Not running Starting usually ~20 s
Leave the lesson

Step 6 of 6 · summary · 2 min

Summary

  • An SLI is the share of good events among all events; for a site, the share of good answers among all requests.
  • Bad answers are 5xx; with 2xx, 3xx and 4xx answers the service coped, and a 404 is the client's mistake.
  • An SLI is named together with its window. The window is selected by the request's unix time $msec and a boundary from date -d '1 hour ago' +%s.
  • The p95 is the line "95 %, rounded up" in the sorted answer times; the mean hides the tail of slow answers.
  • curl at the start of the lesson checked one fast request to the home page. Over the last hour the customers got about 7 % of 5xx answers — even more on /api/ — and one in twenty waited about two seconds or longer for an answer: only the log shows this.
Ubuntu 24.04
Loading…