SLA reports

An SLA report answers “did we hit our uptime commitment last month?” for each UTC calendar month, per monitor, with a clear met or breached verdict.

Find it on the SLA tab of any HTTP monitor, or pull the whole account at once through the API.

Read one

# One monitor
curl -G "https://api.siteqwality.com/http/job/$JOB_ID/sla" \
  -H "Authorization: Bearer $SITEQWALITY_API_KEY" \
  --data-urlencode "months=6"

# Every monitor
curl -G "https://api.siteqwality.com/monitors/sla" \
  -H "Authorization: Bearer $SITEQWALITY_API_KEY" \
  --data-urlencode "months=6"

months defaults to 3 and is clamped to the range 1 to 24. Out-of-range values are clamped rather than rejected, so months=99 quietly gives you 24. The dashboard offers 3, 6 and 12, defaulting to 6.

Each month returns:

Field	Meaning
`month`	`YYYY-MM`, newest first.
`uptime_percentage`	Uptime for the month, or `null` with no data.
`days_with_data`	How many days actually contributed.
`target`	The target this month was judged against.
`attained`	`true`, `false`, or `null` with no data.

The account-wide report adds breached_months per monitor, counting months where attained is false. Months with no data are not counted as breaches.

Where the target comes from

`target_source`	Meaning
`status_page`	Taken from an SLA target you set on a status page.
`default`	No target configured, so the default 99.9% applies.

The target is configured on the monitor’s row in a status page, not on the monitor itself. If a monitor appears on several status pages with different targets, the highest (strictest) one wins.

How the number is calculated

This is the part worth reading before you put a figure in front of a customer.

Month uptime is the unweighted average of that month’s daily uptime percentages. Every day with data counts equally, whatever its traffic. A day with a handful of checks moves the monthly figure exactly as much as a fully sampled day, so a partial day at the start of monitoring, or a day when a monitor was paused for most of it, can pull the month around more than you would expect. Check days_with_data when a number looks surprising.

It is not a request-weighted or minutes-of-downtime figure.

Other rules:

Days with no data are excluded, not treated as 0% or 100%.
A month with no data at all returns uptime_percentage: null and attained: null rather than a breach.
Attainment is inclusive. Exactly 99.900% against a 99.9% target is attained.
Months are UTC calendar months, so the newest month is always month-to-date.

Limitations

HTTP monitors only. Browser, cron, DNS and TLS monitors have no SLA report.
JSON only. There is no SLA CSV or PDF export. The monitor uptime CSV is the closest thing and uses the same daily-average convention.
No scheduled or emailed reports. Reports are pull-only.