Scan a site
ABCDEFGHIJKLMNOPQRSTUVWXYZ
Monitoring

Uptime & SLA

Published: · Updated:

Uptime (availability) is the percentage of time a service is up and responding correctly. An SLA (Service Level Agreement) is a contract in which a provider guarantees a certain availability level and commits to compensation (e.g. credits) when it fails to meet it.

In engineering practice, popularized by Google SRE, three concepts are distinguished: SLI (an indicator, e.g. the share of successful requests), SLO (an internal target for the indicator) and SLA (an external agreement with consequences). The SLO should be stricter than the SLA, and the gap between 100% and the SLO forms the error budget.

What the "nines" mean

  • 99% (two nines) — up to ~3.65 days of downtime per year.
  • 99.9% (three nines) — up to ~8.77 hours per year.
  • 99.95% — up to ~4.38 hours per year.
  • 99.99% (four nines) — up to ~52.6 minutes per year.

How to measure honestly

Availability measured by pinging from a single location inflates the result: a service may respond, but with errors or only in some regions. Reliable measurement relies on probes from multiple locations, verifying response content (not just a 200 status), and complementing synthetic monitoring with data from real users.

Example

A 99.9% SLA gives a budget of ~8.77 h of downtime per year. After a 4 h outage in Q1, ~4.77 h remain for the rest of the year — the error budget in practice.

Error budget for a 99.9% SLA
Example
Annual downtime budget:    ~8 h 46 min (99.9%)
Q1 outage:                   4 h 00 min
Remaining budget (Q2–Q4):    ~4 h 46 min

Sources & further reading

Check your site quality in Insight

See public signals, priorities and data sources in one report.

No credit card Report with date and sources No account required