BACKUP Wait Type in SQL Server

Part of the SQL Server Wait Types Library.
Related deep dive: BACKUPIO and BACKUPBUFFER Wait Types.

---

BACKUP is recorded when a thread involved in backup processing waits to synchronise with the other threads of the same backup. Backups run as a multi-threaded pipeline (readers, writers, a coordinator), and the plain BACKUP wait is the general coordination between them, alongside its more specific siblings BACKUPIO (waiting on the device), BACKUPBUFFER (waiting for a free buffer), and BACKUPTHREAD (waiting for another thread to finish).

Its accumulation tracks how much backup work the instance does.

Is It a Problem?

Not in itself; it has not been a contention point, and time here is proportional to backup activity and duration. The actionable questions live one level up: are backup durations trending the wrong way, and which specific resource is the pipeline waiting on when they do? The sibling waits answer that better than this general one.

Treat plain BACKUP as the backup subsystem’s presence in the wait stats, and diagnose with the specific waits and duration history.

Common Causes

• Normal backup operations, with wait time scaling with backup size and duration.
• Slow spots elsewhere in the pipeline (device, buffers) leaving coordinating threads waiting longer.

What To Do

1. Track durations from msdb.dbo.backupset; trends are the real signal.
2. When durations degrade, split the diagnosis: BACKUPIO up means the destination; BACKUPBUFFER up means buffer starvation (consider BUFFERCOUNT); source disk latency means the reading side.
3. Apply the standard levers: compression, striping to multiple files, and off-peak scheduling.

How To See It

Rank it against everything else with Get-WaitStatistics, and read the whole backup wait family together with your duration history.

---