Extended Events are hard. You have to know the right one to use, pick the right data to collect, and then you’re at the mercy of an unforgiving GUI.
This is my attempt to make them easier and more approachable for the average admin to troubleshoot common scenarios.
- Query performance
- Wait stats
There are lots of great scripts out there that can help you analyze data in your SQL Server, looking at the plan cache, index metrics, aggregate wait stats, etc. The problem is that they can only look at what has already happened, or the before-and-after deltas that a different workload might generate. This proc tightens the net, and helps you catch and log all the things that caused those metrics to spike up.
It’s hard to catch all this stuff just hoping you hit F5 at the right time, and it’s often difficult to get a monitoring tool approved and set up quickly to fill in a lot of these blanks for you.
With sp_HumanEvents, you have two options: Run a quick Extended Events trace to collect data about what’s currently happening on a server, or set up permanent Extended Events sessions and log their data off to permanent tables.
But first, head to my GitHub repo to download the script.
Sampling Live Data
Here are some example calls to get you started.
To capture all types of “completed” queries that have run for at least one second, for 20 seconds, from a specific database
EXEC dbo.sp_HumanEvents @event_type = 'query', @query_duration_ms = 1000, @seconds_sample = 20, @database_name = 'YourMom';
Maybe you want to filter out queries that have asked for a bit of memory:
EXEC dbo.sp_HumanEvents @event_type = 'query', @query_duration_ms = 1000, @seconds_sample = 20, @requested_memory_mb = 1024;
Or maybe you want to find unparameterized queries from a poorly written app that constructs strings in ugly ways, but it generates a lot of queries so you only want data on about a third of them.
EXEC dbo.sp_HumanEvents @event_type = 'compilations', @client_app_name = N'GL00SNIFЯ', @session_id = 'sample', @sample_divisor = 3;
Perhaps you think queries recompiling are the cause of your problems! Heck, they might be. Have you tried removing recompile hints? 😁
EXEC dbo.sp_HumanEvents @event_type = 'recompilations', @seconds_sample = 30;
Look, blocking is annoying. Just turn on RCSI, you goblin. Unless you’re not allowed to.
EXEC dbo.sp_HumanEvents @event_type = 'blocking', @seconds_sample = 60, @blocking_duration_ms = 5000;
If you want to track wait stats, this’ll work pretty well. Keep in mind “all” is a focused list of “interesting” waits to queries, not every wait stat.
EXEC dbo.sp_HumanEvents @event_type = 'waits', @wait_duration_ms = 10, @seconds_sample = 100, @wait_type = N'all';
Note that THREADPOOL is SOS_WORKER in xe-land. why? I dunno.
EXEC dbo.sp_HumanEvents @event_type = 'waits', @wait_duration_ms = 10, @seconds_sample = 100, @wait_type = N'SOS_WORKER,RESOURCE_SEMAPHORE';
For some event types that allow you to set a minimum duration, I’ve set a default minimum to try to avoid you introducing a lot of observer overhead to the server. If you understand the potential danger here, or you’re just trying to test things, you need to use the
@gimme_danger parameter. You would also use this if you wanted to set an impermanent session to run for longer than 10 minutes.
For example, if you run this command:
EXEC sp_HumanEvents @event_type = N'query', @query_duration_ms = 1;
You’ll see this message in the output:
Checking query duration filter
You chose a really dangerous value for @query_duration
If you really want that, please set @gimme_danger = 1, and re-run
Setting @query_duration to 500
You need to use this command instead:
EXEC sp_HumanEvents @event_type = N'query', @query_duration_ms = 1, @gimme_danger = 1;
Logging Data To Tables
First, you need to set up permanent sessions to collect data. You can use commands like these to do that, but I urge you to add some filters like above to cut down on the data collected. On busy servers, over-collection can cause performance issues.
EXEC sp_HumanEvents @event_type = N'compiles', @keep_alive = 1; EXEC sp_HumanEvents @event_type = N'recompiles', @keep_alive = 1; EXEC sp_HumanEvents @event_type = N'query', @keep_alive = 1; EXEC sp_HumanEvents @event_type = N'waits', @keep_alive = 1; EXEC sp_HumanEvents @event_type = N'blocking', @keep_alive = 1;
Once your sessions are set up, this is the command to tell sp_HumanEvents which database and schema to log data to. Table names are created internally, so don’t worry about those.
EXEC sp_HumanEvents @output_database_name = N'YourDatabase', @output_schema_name = N'dbo';
Ideally, you’ll stick this in an Agent Job, so you don’t need to rely on an SSMS window being open all the time. The job creation code linked is set to check in every 10 seconds, in case of errors. Internally, this will run in its own loop with a WAITFOR of 5 seconds to flush data out.
Part of what gets installed when you log data to tables are some views in the same database.
You can check in on them like this:
/*Queries*/ SELECT TOP 1000 * FROM dbo.HumanEvents_Queries; /*Waits*/ SELECT TOP 1000 * FROM dbo.HumanEvents_WaitsByQueryAndDatabase; SELECT TOP 1000 * FROM dbo.HumanEvents_WaitsByDatabase; SELECT TOP 1000 * FROM dbo.HumanEvents_WaitsTotal; /*Blocking*/ SELECT TOP 1000 * FROM dbo.HumanEvents_Blocking; /*Compiles, only on newer versions of SQL Server*/ SELECT TOP 1000 * FROM dbo.HumanEvents_CompilesByDatabaseAndObject; SELECT TOP 1000 * FROM dbo.HumanEvents_CompilesByQuery; SELECT TOP 1000 * FROM dbo.HumanEvents_CompilesByDuration; /*Otherwise*/ SELECT TOP 1000 * FROM dbo.HumanEvents_Compiles_Legacy; /*Parameterization data, if available (comes along with compiles)*/ SELECT TOP 1000 * FROM dbo.HumanEvents_Parameterization; /*Recompiles, only on newer versions of SQL Server*/ SELECT TOP 1000 * FROM dbo.HumanEvents_RecompilesByDatabaseAndObject; SELECT TOP 1000 * FROM dbo.HumanEvents_RecompilesByQuery; SELECT TOP 1000 * FROM dbo.HumanEvents_RecompilesByDuration; /*Otherwise*/ SELECT TOP 1000 * FROM dbo.HumanEvents_Recompiles_Legacy;
If you run into any problems, or would like to discuss adding features, please use my GitHub repo.
Thanks for reading!