Pick the target
Set a target of at least 90 and under 100 percent. For latency, set the threshold in milliseconds. Presets cover the usual choices.
Platform
Set a target, then watch the error budget.
An SLO tracks availability or latency against a target you choose. It shows how much error budget is left and how fast it is being used.
What you get
See the budget going
Each SLO shows attainment, the error budget left and the burn rate over one hour and six hours, and flags the ones at risk.
Says when it cannot tell
With fewer than 10 requests there is no verdict. Thin traffic is marked low confidence instead of being shown as meeting the target.
How it works
Choose a service and an indicator
Availability counts failed requests. Latency counts requests slower than your threshold.
Requests are counted at the entry
The same entry spans Services uses, with health probes left out.
Read the budget
The window defaults to 30 days and is capped at the retention of your plan.
Limits worth knowing
- Burn rate is reported, not alerted on. No alert fires on burn rate yet. Pair an SLO with an alert rule if you need a page.
- The first 50 SLOs are measured. Beyond that, the rest wait.
Turn a promise into a number.
Create your first SLO in a trial workspace.