Documentation › Alert and act
Service level objectives (SLOs)
A numeric objective, an error budget, and an alert when it melts too fast.
Two ways to measure
- From a monitor: the share of time it was OK. A warning or no data count as OK, an alert as bad. Time before the monitor was created is not counted.
- From two queries: the ratio of "good" to "total", for example requests without error over all requests. With no request over the window, the SLI is 100%.
Settings: target (for example 99.9%), optional warning threshold, window of 7, 30 or 90 days.
What HYPHA computes
- SLI: the value measured over the window.
- Remaining error budget: (SLI − target) ÷ (100 − target), as a percentage.
- Burn rate over 1 hour and 6 hours: 1 means the budget runs out exactly at the end of the window.
- State: good, warning, objective missed.
An SLO page plots the cumulative SLI with the target line. The slo dashboard widget and the "SLO" report reuse these values.
Alerting
An slo monitor alerts when the remaining budget falls below a threshold, or when the burn rate exceeds a threshold over 1 h or 6 h. An SLO can also be a component of a status page: a missed objective shows there as "partial outage".
Updated: 19 November 2026