Alerts

PingPane's alert philosophy is described in the quiet hours. This page is the manual that sits behind that philosophy.

When we send an alert

PingPane sends two emails per incident: one when the incident opens, and one when it closes.

An incident opens when a monitor's most recent two checks have both failed. The clock on the incident is stamped with the time of the first failure, not the time of the second.

An incident closes when a monitor's most recent single successful check lands after an open incident. The clock on the close is stamped with that check's time.

When we don't send an alert

• Single failed check. One failure is weather. We wait for the second before firing.
• Flapping. If a monitor has bounced between up and down four times in the last ten minutes, we hold the alert for 60 seconds to see if it settles.
• Unknown state. If we cannot reach our own edge to run the check — a problem on our side — we do not send "down" alerts, because we do not actually know whether your service is up.

Delivery channels

Today: email. One address per account, configurable from Settings → Profile. We are intentionally narrow on channels — shipping SMS and Slack without shipping them well would produce exactly the noise we're trying to avoid.

On the roadmap:

1. Webhooks (signed POST to a URL you own)
2. Slack (via incoming webhook, not an OAuth app)
3. SMS (Twilio, Pro only)

Deliverability

Our mail is sent through Postmark with SPF, DKIM, and DMARC aligned on pingpane.com. If an alert is critical, verify that our domain is on your allow-list. The one place outage emails get lost is in overzealous spam filters.

From: alerts@pingpane.com
Subject: [DOWN] Marketing site (example.com) — started 14:22 UTC

Quiet mode

There is no "quiet hours" setting. The reasoning: if an outage at 3am doesn't matter enough to wake you up, it doesn't matter enough to page you at 9am the next morning either. If you don't want to be woken up, pause the monitor.