Track SLOs and error budgets

Set availability or latency objectives per monitor over 7 to 30 day rolling windows, track error budget and burn rate, and get budget alerts. Team plan.

Updated By the Uptime Tracker team
  1. Open SLOsGo to Reports → SLOs.
  2. Create an SLOChoose "New SLO", enter a Name and pick the Monitor.
  3. Pick the kindChoose Availability (share of time up) or Latency (share of checks faster than a threshold).
  4. Set target and windowEnter the Target (%), for example 99.9, choose a Threshold for latency, and a Rolling window of 7, 14, 28 or 30 days.
  5. Keep budget alerts onLeave "Send budget alerts" ticked to be notified at 50% and 25% budget left, when it is exhausted and on a fast burn.
  6. SaveChoose "Save SLO". The table shows attainment, error budget and burn rate.

A service level objective (SLO) says how reliable a monitor should be over a rolling window, for example "99.9% available over 30 days". The error budget is the downtime, or the number of slow checks, that the objective still allows. SLOs are a Team plan feature.

Create an SLO

  1. Go to Reports → SLOs.
  2. Choose New SLO. The button is disabled until the workspace has at least one monitor.
  3. Enter a Name, such as "Checkout API availability", and choose the Monitor.
  4. Pick the Kind:
    • Availability — share of time the monitor is up.
    • Latency — share of checks faster than a threshold. Choose the Threshold, from 50 ms to 30 s; checks at or under it count as good.
  5. Enter the Target (%), between 0 and 100 (exclusive), for example 99.9.
  6. Choose the Rolling window: 7, 14, 28 or 30 days.
  7. Leave Send budget alerts: 50% and 25% left, exhausted, and fast burn ticked (the default), or untick it for an SLO you only want to watch on screen.
  8. Choose Save SLO.

Read the table

  • Attained: the share of good time (availability) or fast checks (latency) in the window so far.
  • Error budget: how much of the allowance is left, with a bar and "used X of Y". The badge reads Within budget, Budget running low (under 25% left), Error budget exhausted (the objective is already missed for this window) or Insufficient data.
  • Burn rate, for the Last hour and the Last 6 h: how fast the budget is being spent compared with the sustainable pace, the pace that would use exactly the whole budget by the end of the window. 1× is sustainable; 3× means the budget would run out in a third of the window. Each rate comes with a plain sentence, for example "Burning budget 3.2× faster than sustainable — at this rate the budget runs out in about 4 days." or "Within budget (0.4× the sustainable rate)." The last-hour rate reacts quickly; the 6-hour rate shows whether a problem is sustained.
  • Coverage: the share of the window that has check results.

A useful rule of thumb: a 99.9% availability objective over 30 days allows about 43 minutes of downtime; 99.5% allows about 3.6 hours.

How SLOs are calculated

  • Planned maintenance does not count against availability objectives.
  • Below 95% measurement coverage, attainment is not claimed and the SLO shows Insufficient data rather than a guess. New monitors and monitors with long pauses show this until enough data exists.

Budget alerts

Each SLO with Send budget alerts ticked (the default) notifies your destinations when its error budget gets into trouble. SLOs are evaluated once an hour, after each hour closes. There are four alerts:

AlertEventWhen
SLO BUDGET BELOW 50%slo.budget_lowHalf of the window's error budget is used
SLO BUDGET BELOW 25%slo.budget_lowOnly a quarter is left
SLO BUDGET EXHAUSTEDslo.budget_exhaustedThe budget is used up: the objective is missed for this window
SLO FAST BURNslo.fast_burnThe burn rate is at least 14.4× over the last hour and at least 6× over the last 6 hours

A 14.4× burn would use a 30-day budget in about two days, and requiring the 6-hour rate too filters out a single bad minute. The messages say how much budget is left, the objective and monitor, and the burn-rate sentence.

  • Once per threshold: each alert is sent once per SLO and window. It is re-armed only after the SLO recovers above that level, so a budget hovering around 25% does not page you every hour. The table shows which alerts are active, for example Below 25% alert sent.
  • Routing: SLO alerts go through routing rules like outage alerts, so active hours and destinations apply. A rule without an event filter also receives them, while a rule that picks categories never does (SLO alerts have no category). Without rules, every verified, enabled shared destination receives them.
  • Webhooks receive them as slo.budget_low, slo.budget_exhausted and slo.fast_burn events; see webhooks.
  • Turn them off per SLO with Edit: untick Send budget alerts. The table then shows Budget alerts off. Changing an SLO's objective resets its alert state.
  • Budget alerts are sent only while the workspace is on Team (or the Team trial).

Edit or delete

Use Edit or Delete in the table. Deleting stops the error budget tracking; the monitor and its data are kept. If the workspace leaves the Team plan, existing SLOs stay listed but cannot be edited or created.

FAQ

Frequently asked questions

Which plan includes SLOs?

The Team plan, including the Team trial.

Does maintenance use up my error budget?

No. Time inside maintenance windows is excluded from availability objectives.

Why does my SLO say insufficient data?

Less than 95% of the window has check results, for example because the monitor is new or was paused. Attainment is shown once coverage is high enough.

Can an SLO alert me when the budget runs low?

Yes. Budget alerts are on by default: at 50% and 25% budget left, when the budget is exhausted, and on a fast burn (14.4× over the last hour and 6× over 6 hours). Each is sent once and again only after the SLO recovers. Turn them off per SLO with Edit.

Where are SLO alerts sent?

Through your routing rules, like outage alerts. Rules without an event filter receive them too; rules that pick categories do not. Without rules, every verified shared destination receives them.

How often are SLOs checked for alerts?

Once an hour, after each hour closes.

Uptime Tracker

Start monitoring in under five minutes

Start on the free plan — commercial use allowed. No credit card required.

  • Free forever plan
  • No credit card
  • Cancel anytime