Skip to content
BackupSentinel
Start free trial
Documentation menu

PSA ticketing

One ticket per incident in HaloPSA, Autotask, Atera, Syncro or SuperOps, updated and closed as the backup recovers.

Last updated

PSA ticketing turns BackupSentinel incidents into service tickets in HaloPSA, Autotask, Atera, SyncroMSP or SuperOps, with ConnectWise PSA coming soon. A backup that fails or goes missing opens one ticket; further failures update that ticket; when the backup recovers, BackupSentinel adds a recovery note and can close it. Your technicians work the ticket where they already work.

PSA ticketing is part of the Growth plan (one PSA) and the Pro plan (up to three at once, plus per-client routing), and of the free trial.

Setup guides: HaloPSA · Autotask · ConnectWise PSA · Atera · SyncroMSP · SuperOps.

How it works

BackupSentinel stays the source of truth for the incident. Tickets follow it; changes made to a ticket in the PSA are not read back.

In BackupSentinelIn the PSA
A backup fails, goes missing or fails an integrity check (an alert opens)One ticket is created, if that type is ticketed and, for failures, once the job has failed the chosen number of times in a row.
It fails againA note on the same ticket: "failed again, 3 failures in a row". Never a second ticket.
A warning turns into a failureThe ticket's priority is raised to the failure priority, with a note.
Someone acknowledges the alertA note: acknowledged by whom and when.
The backup recoversA recovery note, then the ticket is set to your closed status.
Someone resolves the alert by handA note with their resolution note, then the closed status.
Someone reopens a resolved alertThe ticket is set to your reopen status with a note. If you chose "create a new ticket", or no reopen status is set, a new ticket is created that refers to the closed one.

Each ticket carries the incident's reference, such as BS-1A2B3C4D, in its title. If a request to the PSA fails half-way, BackupSentinel looks for a ticket with that reference before creating one, so a retry does not make a duplicate. The same lookup runs when the incident is resolved while a ticket was being created, so that ticket is still closed.

Ticket content

[BackupSentinel] Backup failed — Pharos Advisory — Microsoft 365 (BS-1A2B3C4D)

Client: Pharos Advisory
Backup job: Microsoft 365
Product: Synology DSM
Status: Failed
First seen: 17 Oct 2026, 02:04 (Europe/Luxembourg)
Failures in a row: 3
Last successful backup: 16 Oct 2026, 02:01 (Europe/Luxembourg)

Summary:
Microsoft 365 backup failed

Likely cause (AI suggestion, high confidence; check before acting):
The Microsoft 365 authorisation expired.

Suggested checks:
- Re-authorise the tenant
- Retry the backup

BackupSentinel incident BS-1A2B3C4D: https://app.backupsentinel.io/dashboard/jobs/…

The AI part appears only when an explanation exists and Include the AI explanation is on. Tickets never contain the report email itself, its sender, ingest addresses or AI prompts.

Set it up

In Settings → Integrations → PSA ticketing, choose your PSA and follow the steps:

  1. Connect. Enter the API credentials (each guide lists exactly what to create in the PSA). BackupSentinel tests them at once. They are stored encrypted and never shown again; secret fields left empty keep the saved value.
  2. Where tickets go. Choose the queue, board or team, the ticket type, the status for new tickets, the status when the backup recovers, the status when an incident is reopened, and a priority for failures and for warnings. Every list comes from your PSA.
  3. Ticket rules.
    • Which incidents create tickets. The defaults are Failed on, Missing on, Integrity check on and Warning off.
    • After how many failures in a row: the first, 2 or 3. A higher number avoids tickets for a single failure that the next run fixes.
    • Which notes to add.
    • Whether to close the ticket on recovery.
    • Whether to include the AI explanation.
    • Whether notes are internal.
  4. Map clients. Each BackupSentinel client needs its PSA company (see below).
  5. Switch on. BackupSentinel lists anything still missing before it lets you switch on. Incidents that were already open get a ticket when they next fail again, escalate or reopen; acknowledging or resolving them does not create one.

Recent ticket updates shows every attempt, its result and errors, with Retry for a failed one.

Client mapping

A ticket must belong to the right customer, so every client is mapped to a company, account or customer in the PSA.

  • Match by name maps a client automatically only when exactly one PSA company has the same name, ignoring case, accents and punctuation.
  • Close names are suggested, never applied on their own. For example, "Pharos Advisory" is suggested for "Pharos Advisory S.à r.l.", and you confirm it with one click.
  • Map and Change search the PSA's companies so you can choose by hand. Remove clears a mapping.
  • Mapping changes are recorded in the activity log.

On Pro, each client can also send its tickets to its own queue or board. That's per-client routing.

Integration health

The top of Settings → Integrations lists every integration, problems first. For example: "Authentication failed. Reconnect below", "3 ticket updates failed in 24 h", or "Connected · 4 open tickets".

On a backup's alert in Needs attention, the ticket appears as "HaloPSA ticket #12345 →" with a link into the PSA.

Troubleshooting

What you seeWhat to do
"rejected the credentials (HTTP 401)" or "Connection failing"The secret was changed, the API user was disabled, or its permissions are too narrow. Choose Change credentials and enter them again; check the permissions in the setup guide.
"… is not mapped to a … company"A client without a mapping had an incident. Map it in step 4, then Retry the failed update.
"refused the request (HTTP 400)" when creating a ticketUsually a queue, board, status or type that no longer exists, or a field the PSA now requires. Choose the mappings again and save.
"is limiting requests (HTTP 429)"The PSA's API limit is reached, often shared with your other tools. BackupSentinel waits and retries by itself (1 minute, 5, 30, then 2 hours).
"is unavailable (HTTP 5xx)" or a timeoutThe PSA is down or slow. Retries are automatic; nothing is lost.
A ticket was not created for an incidentCheck the ticket rules: the type may not be ticketed, or the job has not yet failed the chosen number of times. The skipped update in Recent ticket updates says which.

Limits

  • One direction only: BackupSentinel updates tickets; closing or changing a ticket in the PSA does not change the alert.
  • Ticket updates are sent from BackupSentinel's cloud servers, whose addresses change.
  • An on-premises PSA must be reachable from the internet over HTTPS (ports 443 or 8443).
  • Updates of one incident are sent in order, one at a time per PSA.