Reliability and SRE

Load balancing — copies behind one address and health checks

A load balancer takes requests on one address, spreads them across copies of a service and drops sick ones.

SLI and SLO — the share of good requests and the error budget

SLI: the measured share of good requests over a window. SLO: its target. The gap to 100 % is the error budget.

Incident response — the first minutes, runbook and timeline

An incident is an outage users can see: restore the service by the runbook first, look for the cause later.

Postmortem — a blameless outage review

A postmortem is a written outage review: timeline, impact, root cause and action items, so the outage does not repeat.

keepalived — a floating address over VRRP

keepalived holds a floating address (VIP) on one machine and moves it to the standby over VRRP.