SLA (Service Level Agreement)
Also known as: service level agreement
An SLA is a contract with a customer that specifies a service level and the consequence of missing it — usually a service credit. It is a commercial instrument, not an engineering target, and it should always be looser than the internal SLO so the team detects a problem before the customer invoices for it.
Last reviewed · Part of the Architecture Glossary
In practice
The relationship that should hold:
SLA (99.5%) < SLO (99.9%) <= measured performance
contract internal realityThe gap between SLA and SLO is deliberate margin. If they are equal, the first breach of the internal target is simultaneously a contractual event — no room to notice, escalate and fix before money changes hands.
Contract details engineers should read before they are signed:
- Exclusions. Scheduled maintenance, force majeure, customer-caused errors, beta features. These usually make the effective number very different from the headline.
- Measurement authority. Whose telemetry counts, and at what point in the path.
- Remedy cap. Almost always a service credit bounded by the monthly fee — which is why an SLA is not a reliability strategy.
When it matters
Enterprise sales, vendor selection, and any procurement conversation where a number is about to be promised on the team's behalf.
Common mistake
Letting sales quote "99.99% uptime" with no SLO behind it and no measurement in place. The commitment then exists only in the contract, and the first time anyone computes it is during a dispute.
See also
- SLO (Service Level Objective)An SLO is a target value for an SLI over a window — for example, 99.9% of requests succeed over 28 days.
- SLI (Service Level Indicator)An SLI is a quantitative measure of one aspect of service behaviour, expressed as the ratio of good events to valid events — successful requests over total requests, say.
- Error BudgetAn error budget is the amount of unreliability an SLO permits — the complement of the target.