One ticket per incident, in the PSA you already work in.
A failed or missing backup opens a ticket for the right client. Further failures update it, and the recovery closes it. Your technicians never copy an alert into the PSA by hand.
The ticket follows the incident.
BackupSentinel stays the source of truth. Each incident has one ticket, linked from the alert in Needs attention, and every change to the alert reaches the ticket as a note or a status. A retry after a timeout finds the ticket it already created by the incident reference in its title, so a slow PSA never ends up with two.
You choose which incidents open tickets (failures, missing backups, integrity checks, warnings), whether a failure needs one, two or three runs in a row first, which notes are added and whether they stay internal.
- A backup fails or goes missing
- One ticket opens for the client's company, with the job, the product, the last good backup and, when there is one, the AI explanation's likely cause.
- It fails again
- A note on the same ticket: "failed again, 3 failures in a row". Never a second ticket.
- A warning becomes a failure
- The ticket's priority is raised to the failure priority, with a note.
- Someone acknowledges the alert
- A note says who acknowledged it and when.
- The backup recovers
- A recovery note, then your closed status. Or only the note, if you prefer to close tickets yourself.
- The alert is reopened
- The ticket goes back to your reopen status, or a new ticket refers to the closed one: your choice.
Your PSA, set up from its own lists.
Connect with an API key or application, then choose the queue, ticket type, statuses and priorities from lists read from your PSA. Clients are matched to PSA companies by exact name; near matches are suggestions, and a name that fits two companies is never mapped on its own.
- HaloPSATickets by API application, with team, type and statuses from your HaloPSA.Setup guide →
- Autotask PSAAutotask PSA (Datto, Kaseya): queue, issue type and note type from your picklists.Setup guide →
- ConnectWiseComing soonConnectWise PSA: board, type, subtype and the board's own statuses.Setup guide →
- AteraBetaTickets for the customer's contact, with technician notes.Setup guide →
- SyncroSyncroMSP with a scoped API token. Nothing is emailed to your customer.Setup guide →
- SuperOpsBetaTickets with source Integration, technician group and private notes.Setup guide →
Signed webhooks for everything else.
Send each alert change to n8n, Make, Zapier, an internal tool or a script. Every request is a JSON POST with the client, the job and the alert, signed with a secret that is shown once when the endpoint is created.
alert.created- A backup failed, warned, went missing or failed an integrity check.
alert.updated- It failed again, or a warning became a failure.
alert.acknowledged- Someone acknowledged the alert.
alert.resolved- The backup recovered, or someone resolved the alert.
alert.reopened- A resolved alert opened again.
backup.failed, backup.warning, backup.missing and backup.recovered are shorter names for the same changes. An endpoint never gets two requests for one change.
POST https://n8n.example.com/webhook/backups X-BackupSentinel-Event: alert.created X-BackupSentinel-Delivery: 6f1c… X-BackupSentinel-Timestamp: 1792289071 X-BackupSentinel-Signature: v1=9b2e… # v1 = HMAC-SHA256(secret, timestamp + "." + body) # reject anything older than 5 minutes
- Retried after 1 minute, 5 minutes, 30 minutes and 2 hours, honouring Retry-After
- The same event id on every retry, so your side can ignore a repeat
- A delivery log with the status, the time taken and the start of each answer
- Send test, retry, and an endpoint that pauses itself after 25 failed deliveries in a row
Your PSA credentials stay ours to protect.
- Encrypted, never shown again
- PSA credentials and webhook secrets are encrypted when saved. Settings shows a hint such as “Client secret ••••”, never the value, and changing the PSA’s address asks for the secret again.
- Only public addresses
- Webhook and PSA addresses must use HTTPS on the public internet. Private networks, loopback and cloud metadata addresses are refused when saved and again for every request, and redirects are not followed.
- Facts, not raw emails
- Tickets and webhooks carry the client, the job, the status and the alert’s summary line. Never the report email itself, its attachments, headers or credentials.
- Owners and admins
- Only owners and admins connect a PSA or add an endpoint, and every change is in the activity log. Members and viewers see the health of each connection and the delivery log.
On which plan.
- Starter
- 1 webhook endpoint. No PSA ticketing.
- Growth
- 5 webhook endpoints and 1 PSA.
- Pro
- 25 webhook endpoints, up to 3 PSAs at once, and a different queue or team per client.
- Free trial
- Everything on Growth for 14 days.
What it does not do.
- Tickets are not read back: closing a ticket in the PSA does not resolve the alert in BackupSentinel.
- One PSA ticket per incident and connection. It does not create tickets for healthy backups or for reports.
- Atera and SuperOps are in beta: they follow the vendors' published APIs and have not been checked against every setup.
- ConnectWise PSA is coming soon. NinjaOne, Kaseya BMS and other service desks are not connected yet; a webhook reaches anything that takes an HTTP request.
Let the PSA hear about failed backups first.
Connect your PSA, map your clients, and the next failed backup opens a ticket on its own. The trial runs 14 days with no card.