Peer Confirmation

exit1 runs every check from a primary region (Frankfurt or Boston). When a check fails, the primary region asks the peer region to probe the same target. Your check is only marked DOWN if both regions agree.

This eliminates false alerts caused by transient network issues that affect only one region — the most common source of noise in single-region uptime monitoring.

How it works

1. The primary region runs your check on its normal schedule.
2. If the check succeeds, nothing else happens — the peer is never consulted on healthy probes.
3. If the check fails, the primary asks the peer region to probe the same URL/host.
4. If the peer also reports DOWN, the failure counts toward your alert threshold (3 consecutive failures within 5 minutes).
5. If the peer reports UP, the failure is ignored — most likely a transient network issue between the primary region and your service.
6. If the peer can’t be reached (timeout, network issue, etc.), the check falls back to the original time-based confirmation. Behavior is identical to before peer confirmation existed.

What you see in alerts

When both regions agree a check is down, your alert email includes a confirmation line:

Confirmed by peer region (vps-us-1): both regions reporting DOWN.

If the peer was unreachable, this line is omitted — the alert still fires (using the temporal confirmation gate), but you know the disagreement was about peer connectivity, not your service.

Plan availability

Peer confirmation is on for every check on every plan, including the free tier. It is positioned as a reliability addon, not a paid upsell. If the peer call fails for any reason, exit1 falls back to time-based confirmation — peer confirmation can only ever make alerts more reliable; it can never make them worse.

When to disable it

Most checks should leave peer confirmation enabled. The exception: endpoints that legitimately respond differently from different geographies, such as:

• A service geo-blocked to EU users (Frankfurt sees 200, Boston sees 403).
• A site behind a WAF that issues a CAPTCHA challenge to non-residential US IPs but not to EU IPs.
• A regional endpoint like eu.api.example.com that intentionally rejects requests from outside its region.

For these checks, primary and peer will permanently disagree, and peer confirmation would suppress every failure forever — meaning you’d never get paged for a real outage.

To disable peer confirmation for a specific check, open the check’s edit form and toggle Disable peer confirmation in the Alert behavior section.

Permanent-disagreement notification

If your check is stuck in disagreement (peer always says UP, primary always says DOWN) for more than 30 minutes continuously, exit1 sends you a one-time low-priority email pointing this out and linking to the disable toggle. We don’t re-send for 24 hours per check.

This catches the failure mode where you have a misconfigured check that would otherwise never alert.

Network setup

Peer confirmation requires both monitoring regions to reach your service. If you allowlist exit1’s IPs, you must allowlist both the Frankfurt and Boston IPs — even if you only selected one region for the check. The peer region needs the same network access as the primary in order to confirm a failure.

See Check Regions for the current list of monitoring IPs.