Docs

The monitor page

What each part of a monitor's page shows and how uptime, latency, and the recent checks are calculated.

This page explains what each part of a monitor's page shows and how the numbers are calculated. The page refreshes itself every 25 to 40 seconds while it is open.

#The toolbar

The top row holds the target (a link for HTTP family and scheduler monitors), Pause or Resume, Refresh, the range picker, and the settings gear that opens Monitor settings.

The range picker offers 1 hour, 6 hours, 24 hours, 7 days, 30 days, and 90 days, and defaults to 24 hours. Ranges longer than your plan keeps are locked: 7 days of history on the free plan, 30 on Pro, 90 on Team.

A banner above the page says when the monitor is paused, under a maintenance window, or has an incident open.

#The header

The monitor's name, its current status with how long it has held it, the interval, and when the last check ran.

#The KPI row

Four numbers summarize the selected range:

  • Uptime, as a percentage to two decimals. Up and slow checks both count as up. Checks skipped during a pause or a maintenance window are left out.
  • Avg, Latest, and Peak latency. Latest is the newest check whatever the range. The other two follow the range.

The word behind the latency changes with the type, because the measured time means something different: response time for HTTP, Keyword, API, and Scheduler monitors, connect time for Port, and resolve time for DNS. A heartbeat has no latency, so its row shows On-time and Missed instead, next to a Cadence card that counts down to the next expected ping and to the moment the monitor would go down.

#The chart

The latency chart plots the range as a line with Average, Fastest, Slowest, p50, p95, p99, and Latest in its legend. Ranges up to 24 hours plot individual checks. Longer ranges plot hourly averages.

#Recent checks

The recent-checks bar shows the last 30 checks as cells, oldest on the left: green for up, amber for slow, red for down, and empty for a slot that was skipped while the monitor was paused or under maintenance. Hover a cell for its time and latency. A tally of up, slow, and down counts sits beside it.

#Type-specific panels

TypePanel
HTTP, Keyword, APIBreakdown of last check: the time to connect (DNS, TCP, and TLS) and the time waiting for the first byte. Connect time rising points at the network or the host. Waiting time rising points at your backend.
HTTP, Keyword, APITLS certificate and Domain cards with the days left on each, when the checks are on. See SSL and domain expiry.
Keyword, APIContent match or JSON assertions, with the last check's verdict rule by rule and what the body contained.
HeartbeatThe ping URL with a curl example, and the Cadence card.
PortThe TCP endpoint and its connect time as signal bars.
DNSResolution: the record type, the query, the answers from the last resolve, when they last changed, and the expected-value verdict.
SchedulerTrigger: the method, the endpoint, how often it fires, and the result of the last run.

#Activity

The bottom of the page is the event log: every confirmed status change with its time and detail, including which content rule failed and what the body contained, plus the actions people took on the monitor, such as pausing, resuming, and edits with a summary of what changed. The page shows the latest 50 events.

#How the data is kept

UptimeX stores each monitor's results as hourly totals plus the sequence of its most recent checks, not as a row per check. Totals are kept for 7 days on the free plan, 30 days on Pro, and 90 days on Team, and the range picker and activity log are capped at that window.

#What's next?