Latency SLO

A latency SLO measures the fraction of checks that returned below a response-time threshold. Applies to http_check monitors only.

Computation

The SLO holds when the fraction of checks under the threshold is at least the target percentile divided by 100:

latency_slo = fraction_of_checks_below_threshold ≥ target_percent / 100

Setting target_percent = 95 and threshold = 500 requires p95 ≤ 500 ms over the rolling window.

Configuration

Open the monitor edit form and scroll to SLO Settings. Latency fields:

latency_threshold_ms — Threshold (ms)
Response-time ceiling in milliseconds. Checks under this value are counted as fast. Tunable per monitor on every plan.
latency_target_pct — Target percentile (%)
Minimum fraction of checks that must complete under the threshold. 95 means 95% of checks must be fast. Business and Founder only; defaults to 95.
latency_excluded_regions — Excluded regions
Per-region toggle. Excluded regions still run checks and contribute to availability; their response times are removed from the latency aggregate and burn-rate math. Business and Founder only.

Latency panel

Current p95
p95 over the full SLO window (default 7 days). Drives the error-budget calculation.
Recent p95 (1 h)
p95 over the last hour. Reacts quickly to recent changes — deploys, traffic spikes, regional network events.

A rising Recent p95 that has not moved Current p95 indicates a recent change that has not yet affected the budget.

Per-region breakdown

Below the main latency card, a grid shows p95 per probe region. Excluded regions are tagged with an excluded badge; their p95 is still displayed for reference. Regions whose p95 exceeds the threshold are highlighted.

Plan tier matrix

CapabilityFreeProBusinessFounder
Latency SLO panel and per-region breakdown
latency_burn detector
latency_threshold_ms
latency_target_pct, rolling window, latency_excluded_regions

Defaults applied to Free and Pro: 95% / 500 ms / 7 days. The threshold can be overridden per monitor on every plan.

See also