Skip to main content

How three-check and multi-region confirmation works

Understand the difference between three monitoring regions and three consecutive failed check cycles before a monitor becomes Down.

Uptime Basics combines two protections before changing an availability monitor to Down:

  1. Regional confirmation: at least two of three monitoring regions must report failure for one check cycle.
  2. Consecutive confirmation: three region-confirmed failed check cycles must occur in a row.

Three regions and three consecutive checks are separate requirements.

Step 1: The primary check runs

The normal scheduled check runs from Canada. If it succeeds, no additional regional confirmation is needed and the monitor's consecutive-failure count resets.

The request uses the monitor's URL, method, timeout, redirect setting, accepted status codes, custom headers, and authentication configuration. A supported HEAD-to-GET fallback occurs before the result is treated as a final primary failure.

Step 2: Additional regions confirm a failure

When the primary request fails, Uptime Basics requests the same availability probe from Eastern United States and Western United States.

The current policy is 2 of 3:

Region results Check-cycle decision
Primary fails; both additional regions fail Region-confirmed failure
Primary fails; one additional region fails and one succeeds Region-confirmed failure because two regions failed
Primary fails; both additional regions succeed Failure not confirmed
Primary fails; only one additional result is available and it succeeds Failure not confirmed
Primary fails; additional probes are unavailable Failure not confirmed without enough failure evidence

An unavailable confirmation probe is not automatically counted as a failed website. The service requires enough executed failure evidence to make the Down decision.

When regional evidence does not confirm the primary failure, the monitor does not add a confirmed failed cycle. Detailed Logs preserve the primary failure and regional outcome so you can still investigate the difference.

Step 3: Consecutive failed cycles are counted

One region-confirmed failed cycle increases the consecutive-failure count by one. The monitor becomes Down only after three consecutive region-confirmed failed cycles.

Any effective successful cycle resets the count to zero. For example:

Confirmed failure -> Confirmed failure -> Success -> Confirmed failure
Count:             1                    2          0          1

This sequence does not open an availability incident because it never reaches three in a row.

Step 4: The incident and Down alert are created

On the third consecutive confirmed failed cycle:

  • The monitor changes to Down.
  • An availability incident opens.
  • Enabled email and SMS channels can queue a Down notification.
  • Incident diagnostics can be requested.

The incident start time is the confirmed Down transition, not the timestamp of the first failed row.

Recovery behavior

After the monitor is Down, one effective successful check cycle changes it back to Up, resolves the incident, and can queue a recovery alert.

Regional confirmation is invoked only when the primary request fails. If the primary request succeeds, that successful result is enough to recover the monitor. If the primary fails but regional evidence prevents a confirmed failure, the cycle is treated as reachable for the monitor decision.

How long confirmation can take

Confirmation time depends on:

  • The configured check interval.
  • The timeout for each request.
  • Regional response time.
  • Scheduling and processing time.
  • Notification delivery after the state change.

A one-minute interval does not guarantee a Down alert exactly three minutes after a problem begins. The outage may start between checks, and each failed cycle must finish its regional evaluation.

Read confirmation in Detailed Logs

The Confirmation column can show:

  • Down confirmed from 2 regions or 3 regions.
  • Primary failure not confirmed; healthy in ... when another region reached the website.
  • Primary failure was not region-confirmed when there was not enough regional failure evidence.

Use this information when your browser works but a monitoring region reports a failure. Regional routing, DNS, firewall, bot protection, and provider behavior can produce different results from different locations.

Related articles

Did this answer your question?

Your response helps improve this Help Center.