Availability SLO

An availability SLO measures the fraction of checks that succeeded. Nines computes your error budget in real time and alerts you when the burn rate threatens your target.

How availability is measured

For every check interval, each selected region probes your monitor target and records a result of up or down. The availability fraction is:

availability = up_checks / total_checks

This ratio is computed over a rolling window (7 days by default) and expressed as a percentage. A value of 99.95% means that 0.05% of check results over the window were down.

Configuring the availability SLO

Open the monitor edit form and scroll to SLO Settings. Two fields control the availability SLO:

  • SLO target (%) (slo_target_percent) — the minimum acceptable availability. Common values: 99.9 (three nines), 99.5, 99.0. Leave blank to inherit the plan default (if any).
  • SLO window (days) (slo_window_days) — the rolling look-back period. 7 days is the default and a good starting point. Shorter windows react faster to recent failures; longer windows smooth out noise.

Reading the availability panel

Once data is available the panel on the monitor detail page shows:

  • Budget remaining — the percentage of your total error budget that has not been consumed. If your target is 99.5% over 7 days, you have 0.5% of 604,800 seconds ≈ 50 min of allowed downtime. If you have used 10 minutes, the budget remaining is approximately 80%.
  • Used / Allowed — raw downtime consumed vs. total allowed, shown in a human-readable format (e.g., "10 min used / 50 min allowed").
  • Burn rate — how fast the budget is being consumed relative to the sustainable rate. A burn rate of 2.0× means you are consuming the budget twice as fast as needed to stay within the SLO.

What happens when data is not yet available

A new monitor or a recently reconfigured monitor may not have enough check history to compute a meaningful error budget. In that case, the panel shows a Collecting data placeholder. Budget figures appear once enough checks have accumulated to fill the configured window.

Interpreting the burn-rate classes

  • Normal — burn rate ≤ 1.0×. You are on track to end the window with budget to spare.
  • Warning — burn rate > 1.0× but budget not yet exhausted. You are trending toward a breach — investigate now.
  • Exhausted — the error budget has been consumed and the SLO target was missed for this window.

How check interval determines SLO precision

Every failed check creates an interval of unknown downtime bounded by the check spacing. If your monitor fires every 5 minutes and one check comes back down, the SLO calculation can't tell whether the outage lasted 1 second or the full 5 minutes — it has to charge the worst case against your error budget. For a 99.9% / 30-day SLO (43.2 minutes total budget), the same incident lands very differently depending on your check interval:

Worst-case budget impact per failed check — 99.9% / 30-day SLO
Check interval Budget per failure Failures to exhaust budget
5 min~11.6%8
1 min~2.3%43
30 sec~1.2%86
10 sec~0.4%259

The check interval available to your monitor depends on your plan. See the Billing & Plans page for per-plan interval options. For the related angle on detection gaps — outages that fall entirely between two checks and are never recorded at all — see What 5-minute check intervals actually miss.

Availability SLO and plan tiers

The availability SLO panel is available on Business and Founder plans. On the Free plan the error-budget section is hidden. Upgrade your plan from Settings → Plan to unlock SLO tracking.