Uptime
Reading the results
Statuses, uptime percentages, the response-time chart and the check history.
Status
| Status | Means |
|---|---|
| UP | The last check succeeded |
| DOWN | The first failed check plus every retry failed (4 in a row by default; set per monitor with Retries before Down) |
| PAUSED | Checking is stopped. Nothing is recorded |
| PENDING | Added recently; not enough checks yet |
Retrying before DOWN stops a single network hiccup being reported as an outage.
Uptime percentages
Shown over 24 hours, 7 days and 30 days: the share of checks that succeeded in that window.
A little perspective on what the numbers mean over 30 days:
| Uptime | Downtime per month |
|---|---|
| 99.9% | about 43 minutes |
| 99.5% | about 3.6 hours |
| 99% | about 7 hours |
| 95% | a day and a half |
A paused monitor records nothing, so pausing during planned maintenance keeps the percentage honest.
Response-time chart
Fourteen days of how long each check took.
- A flat line is a healthy site.
- A steady upward drift over days or weeks usually means a database growing without an index, or a disk filling up.
- Spikes at the same time each day point at a scheduled job — check Scheduled on the server.
- A step change means something changed: a deployment, a config edit, a new neighbour on shared hosting.
Check history
Thirty days of individual checks: when, the result, the status code and the time taken. This is where you go after an outage to answer "how long was it actually down?"
The compact bar strip on the monitor list shows the same data at a glance — each bar is one check, red where it failed.
When a monitor says DOWN but the site looks fine to you
Both can be true.
- You may be looking at a cached copy. Your browser, or a CDN like Cloudflare, can serve a page while the origin server is dead.
- The failure may be intermittent — one request in five. Look at the check history.
- It may be geography. We check from our servers; you check from your office.
- Your expected status or text check may be wrong. A site that legitimately
redirects will fail a strict
200check.
Cross-check with the server itself: open it and look at Services and Logs, or ask the assistant to investigate the URL.