Vigilancia de disponibilidad: alertas cuando un cajero se cae (Pro)

Uptime watch works with machines that send a heartbeat or a feed. It's part of Pro. On Free you still get an email once a machine has been quiet for 2 hours.

What it alerts on

• Stopped reporting: no heartbeat or feed update for your chosen time (15, 30, 60 or 120 minutes). The public listing still turns Offline at 2 hours.
• Faults your machine sends with its heartbeat: bill_validator_jam, cashbox_near_full, dispenser_low, crypto_float_low. Send "faults": [] when they clear. Faults are never shown on the public listing.
• A visitor says it's not working, from the listing's report button.

When the problem clears, you get a "Resolved" message with how long it lasted. A temporarily closed machine never alerts as down.

Where alerts go

• Email to the whole team, or to the teammate on call for that machine (set it in the By machine table).
• Slack: paste an incoming-webhook URL for the channel you want.
• Webhook: we POST JSON to your https URL, signed with X-BitcoinScouts-Signature: sha256=<HMAC-SHA256 of the body with your secret>. Use it for PagerDuty, Opsgenie or your own tools. The secret is shown once when you save; make a new one any time.

Acknowledge, escalate, quiet hours

Every alert has an Acknowledge link that works without signing in. If nobody acknowledges within 30 minutes, it goes to your escalation email. During quiet hours we hold alerts and send the ones still open when quiet hours end.

Weekly report

Every Monday the team gets last week's uptime per machine: hours down, outages, mean time to recover, faults and visitor reports, worst machine first. It's the evidence for a conversation with a host. The same numbers are on the Uptime watch page for 7 or 30 days.

Use Send a test alert after changing settings: it reports which channels worked.