How does Error Budget work in SRE?

Question

Accepted Answer

**Error Budget** — acceptable failure rate calculated based on SLO.

**Calculation:**
```
Error Budget = 1 - SLO
If SLO = 99.9% → Budget = 0.1% = 43.2 min/month
If SLO = 99.95% → Budget = 0.05% = 21.6 min/month
```

**Burn Rate (Prometheus):**
```promql
(
  1 - (
    sum(rate(http_requests_total{status=~"2.."}[1h])) /
    sum(rate(http_requests_total[1h]))
  )
) > (14.4 * 0.001)
```

**When budget is exhausted:**
1. Feature freeze.
2. Only fixes.
3. Mandatory Post-mortem.
4. Improve monitoring.

DevOps: How does Error Budget work in SRE?