Work out cloud sla composite availability instantly with clear inputs, formula shown and shareable results.
Composite availability multiplies across every service in the request path, so six dependencies at 99.95 percent give 99.7 percent — around 26 hours of expected downtime a year. Redundancy on individual services counteracts this by squaring their failure probability, which is why the number of non-redundant dependencies is the most important architectural availability decision.
Composite SLA
redundant service availability = 1 - (1 - SLA)^replicas; composite = SLA^(non-redundant services) x redundant availability^(redundant services)
Because they multiply. Each additional hard dependency subtracts from your achievable availability, which is why graceful degradation matters more than provider SLAs.
No. SLAs are financial commitments with exclusions, not predictions. Observed availability is usually better than the SLA, but the SLA is what you can plan against.