Waitopedia is a comprehensive resource of information about SQL Server waits.

The description shown below is the top answer as voted by the Spotlight community.

The charts are based on 2.1 TB of data collected from 4207 instances uploaded by 323 Spotlight users over an 8 week period.


Category: Idle
SQL Server versions: 2008, 2008 R2, 2012, 2014, 2016, 2017

This is an idle wait and so it can be safely ignored.


2 Answers


Occurs when the Service Broker lazy flusher flushes the in-memory transmission objects to a work table. Information from Microsoft®


For performance reasons Service Broker maintains all dialog state (TO - transmission object) in memory as well as in temporary tables on disk. Every time a TO is updated, it is scheduled to be flushed lazily to the temporary table on disk. Service Broker employs an always alive lazy flusher task to do this job.

This wait type is charged when the TO lazy flusher task is waiting for some TOs to be saved to the temporary tables. The lazy flusher sleeps for 1 second before waiting again for ~1 second for TOs to be saved.

If Service Broker is not used at all, wait_time_ms and waiting_tasks_count for this wait type should be proportional to the duration since instance start up, with avg_wait_time_ms being close to ~1 second. When Service Broker is used heavily these columns should have low values since the lazy flusher will be busy as well. Source: http://blogs.msdn.com/b/sql_service_broker/archive/2008/12/01/service-broker-wait-types.aspx

Very rare

This wait never occurs for most instances

For each of 4207 instances, we ranked BROKER_TO_FLUSH on how frequent it is compared to all other recent waits. The chart shows the total of all rankings.

Prevalence of BROKER_TO_FLUSH across the Spotlight Population

For 80 % of hours with this wait, average wait time is around 1 seconds

For each instance, we found all the recent hours when it had a BROKER_TO_FLUSH wait. We found the average latency for each of those hours.

1152 instances contributed data to this chart

Latency Distribution of BROKER_TO_FLUSH across the Spotlight Population