SQL Server 2012 :: Right Way To Query Extended Events Session?
Jun 4, 2015
I use this code in a utility procedure (for performance testing) but it is really slow.
For example, a session with three events is taking 5 seconds to complete this query:
DECLARE @xml xml=
(
SELECT CAST(xet.target_data AS xml)
FROM sys.dm_xe_session_targets AS xet
JOIN sys.dm_xe_sessions AS xe
ON (xe.address = xet.event_session_address)
WHERE xe.name = @name
);
I wanted to demonstrate something about CXPACKET wait type. For the purpose of this demo I've created a query in AdventureWorks database that uses a parallel query plan, an extended events session that captured the wait statistics for a single session and a query that shows the extended event's data. I ran it and it worked fine. Then I dropped and recreated the event session (to clear the data), in a new window I wrote a transaction that updated the table fallowed by waitfor statement so the first query will be blocked for few seconds and ran the whole thing again. The select statement was blocked as expected (ran for 9 seconds instead on 1 second as it ran without the blocking), but the wait stats that I got were almost identical to the those that I got without blocking the query.
--Query that uses parallel query plan With MyCTE as ( select top 50 * from Sales.SalesOrderHeader) select top 10000 * from Sales.SalesOrderHeader, MyCTE order by newid()
I have created a session in extended events and want to frequently monitor the events that i have filtered .The problem i am facing is i am not able to clear the previous data as we do it in SQL trace. i am able to see all the data that is occupied till now.whenever i start the test i want the current data only to be displayed.The work around is either delete the session and create a new one every-time which i do not want to do .I am able to see the CLEAR DATA FROM TABLE button but it is in Disabled state all the time.
I am trying to use xquery to get event data back from extended events. I am trying to use some sample data from Grant Fritchey but I am getting null records back. Below is the xml - I just want to retrieve a distinct list of the client_hostname and client_app_name.
WITH xEvents AS (SELECT object_name AS xEventName, CAST (event_data AS xml) AS xEventData FROM sys.fn_xe_file_target_read_file ('C:LoginTraceShared_0*.xel', NULL, NULL, NULL)) SELECT distinct top 1000 xEventName, xEventData.value('(/event/data[@action_name=''Client_APP_Name'']/value)[1]','varchar') Client_APP_Name, xEventData.value('(/event/data[@action_name=''Client_Host_Name'']/value)[1]','varchar') Client_Host_Name FROM xEvents
We are planning to convert or change all existing Traces to Extended Events in SQL server 2012. What is the procedure to convert custom traces. We have already created some below custom traces: like this we are planning to convert for all servers.
I've setup a deadlock monitor using extended events like this.
CREATE EVENT SESSION [deadlock] ON SERVER ADD EVENT sqlserver.lock_deadlock( ACTION(package0.process_id,sqlserver.client_app_name,sqlserver.client_hostname, sqlserver.database_id,sqlserver.database_name,sqlserver.plan_handle)
[Code] ....
Deadlock happened couple of days ago. I'm trying to determine the cause of deadlocks. What script should I use to pull that information to see what objects/processes caused deadlock?
I want to gather (and filter by) client ip address in an extended events session. So far, I have the sqlserver.login event added, but I can find no way to get client ip address added to the action or predicate.
I can filter by the pred_source field "session_nt_domain" or "nt_domain" but this is not always populated by all logins and also, this domain only lists machine name, such as FRED, as opposed to the full domain FRED.foo.bar.
Is it possible to gather login client ip address in an extended events session? Surely internal SQL Server processes have access to it because the login audit can populate client ip address.
I have an existing EE setup that captures all failing queries (see code below). The problem is that I also want to somehow capture RPC_starting so that I can see which parameters are passed in whenever a query fails. Is there a way to somehow capture those two events (error_reported & rpc_starting), but only capture rpc_starting when there is actually an error reported?Or maybe just an event on rpc_starting and somehow filter to only capture when there is an error?
Existing error_reported EE code: CREATE EVENT SESSION [what_queries_are_failing] ON SERVER ADD EVENT sqlserver.error_reported ( ACTION(sqlserver.sql_text, sqlserver.tsql_stack, sqlserver.database_id, sqlserver.session_id, package0.collect_system_time, sqlserver.transaction_id, sqlserver.username, sqlserver.client_hostname)
I do have the following Extended Event session on my Dev box;
CREATE EVENT SESSION [sp_showplan] ON SERVER ADD EVENT sqlserver.query_post_execution_showplan(SET collect_database_name=(1) ACTION(sqlserver.plan_handle) WHERE ([package0].[equal_uint64]([object_type],(8272)) AND [sqlserver].[equal_i_sql_unicode_string]([object_name],N'MyStoreProcedure'))) ADD TARGET package0.event_file(SET filename=N'E:DBA_AuditSP_Exec.xel',metadatafile=N'E:DBA_AuditSP_Exec.xem') WITH (MAX_MEMORY=4096 KB,EVENT_RETENTION_MODE=ALLOW_SINGLE_EVENT_LOSS,MAX_DISPATCH_LATENCY=30 SECONDS,MAX_EVENT_SIZE=0 KB,MEMORY_PARTITION_MODE=NONE,TRACK_CAUSALITY=OFF,STARTUP_STATE=OFF) GO
The idea is being able to capture the execution plan when the program invokes the store procedure, regardless of the database.
This works on my Dev box. When I manually trigger "MyStoreProcedure" from database A , the event is saved. The same thing happens when I do that from database B. Ok ... so far, so good.
So I went to the live production environment and setup my Extended Events session. But it's saving nothing. I was able to check that the store procedure was executed on several databases but my extended events session never grabbed the plan.
What could be the reason for this? Memory starvation maybe? Is there something I am doing wrong?
When you view the Extended Events "Watch Live Data" is the duration in milliseconds or microseconds? I'm assuming it's milliseconds, but if you look at the timestamp difference from start to complete it doesn't add up to the duration amount? It looks like it's just the difference between the timestamps?
In the past, I've combined server side traces with Perfmon successfully, which is pretty useful, I know that. I would like to do the same with Extended Events, so I can correlate and analyze the server side as well.
I am wanting to get/filter on all queries and procs that take longer than 2 seconds to run (I'll balance real values later) but I'm not sure which Action out of the XE that I need.
I am using SQL Server 2014 and thought I had used sqlserver.sql_statement_completed.duration > 2000 in a previous version.
I have two extended events sessions running on a server. I do have sql jobs that automatically stop the XE sessions and import the results to working tables. I see two "ALTER EVENT SESSION XXXX ON SERVER STATE = STOP" statements that are being executed for more than 2 days, the wait types are XE_SERVICES_MUTEX and PREEMPTIVE_XE_SESSIONCOMMIT. This is not the first time I see this behavior.I do not want to kill the sessions (I guess the sessions won't die anyway) neither restart the sql service
I know very little about Extended Events, but I know it is supposed to be more powerful than Profiler. The SQL Instance involved is 2008R2. I was asked whether we could capture when a stored proc was called with a certain parameter. So for example, if dbo.usp_mystoredproc is called and is passed a value of '12345a' for @customer, can that be captured? I would want to know the time it was called, the parameter and parameter value. Probably would be interested in the SPID or LoginName as well.
I need to find any stored procedures that have not been used over a certain time period.
I have set up an extended events session to gather sp_statement_starting and _completed.
The trace returns the object_id's of many stored procedures. I then query the sys.procedures and plug in the object_id to return the stored proc name.
Do I have to repeat this process for every different object_id or is there a way I can query the trace results, using the object_id as my search criteria as one query ?
I am setting up extended events more or less just fine, however I am a bit confused as to how to read and load them into a table for querying. In particular the offset part - is there a way to load just a given dates worth in?
I've got the files configured to be 20MB before rolling over, the XE is running all the time.
So if i load in the full file now, say that covers 2.5 days worth, when I load it again tomorrow to get the updated data I'm also reloading today, which is a waste?
I presume I am going about this wrong, but lack an example that really goes into detail of practicalities of loading this data.
I can only do one match at a time -- Like can only do either the sql_statement_(start and end), or sp_statement_(start or end). Is there any way to capture both in the same session? Or since I am adding both the events in the ADD EVENT section, can I query it somehow to get unmatched SP or SQL?
Code:
USE master; GO -- Create the Event Session IF EXISTS(SELECT * FROM sys.server_event_sessions WHERE name='TimedOutSQL')
We know we can use the event lock_deadlock and xml_deadlock_report to capture the deadlock info, however I also want to capture the execution plans for all of the SPIDs in the deadlock graph, how to output the execution plans to the extended events trace results either ? such as if there is an action for execution plan or workaround for it ?If there is no built in action for execution plan , may I know if we can add the customized info to the extended events results file also ? Such as when the deadlock related event happens , then we can run a query to get some info ,then added the info along with other info such as sql_text, dbname etc to the events trace results file either ? The reason is if we also know the execution plans when the deadlock happens, it is useful to turning the query based on the execution plans to reduce deadlock happening .
Hello experts. I have been searching for anything about this but found very little. What are the events logged in SQL Server Error Logs aside from Successful/Failed Login, Backup/Restore/Recover database and, start/init sql server? Can we configure this to log other events, like CREATE or DBCC events for example? If so, how? Thanks a lot.
The 2.0 version of ASPSTATE is slightly different than the 1.1 version in that one table has one additional column and another table uses a different data type and size for the key. The 2.0 version also has a couple additional stored procedures.
We'd like to manage just one session state database if possible so we're trying to figure out if Microsoft supports using the new schema for 1.1 session state access (it seems to work, but our testing has been very light).
Is there any official support line on this? If not, can anyone comment on whether or not you'd expect it to work and why?
I am logging into a SQL instance to run the following query:
DECLARE @ReturnCode int EXECUTE @ReturnCode = [master].dbo.xp_create_subdir N'sharemasterFULL' IF @ReturnCode <> 0 RAISERROR('Error creating directory.', 16, 1)
The share in which the folder is to be created has my account added with full permissions to create files. However this command fails unless I add the SQL Service account user with rights to the folder also.
Is this expected behaviour, is this something specific to extended stored procedures?
I am learning Extended event to capture deadlock which already happened, for this in my SQL SERVER 2012.. I am simulating a deadlock . With [URL]... where given a query to find the deadlock details using extended event here is the code..Retrieve from Extended Events in 2012.
SELECT XEvent.query('(event/data/value/deadlock)[1]') AS DeadlockGraph FROM ( SELECT XEvent.query('.') AS XEvent FROM ( SELECT CAST(target_data AS XML) AS TargetData FROM sys.dm_xe_session_targets st JOIN sys.dm_xe_sessions s
[code]...
This code creates a deadlock but when i run the above Extended events query to get the details of deadlock, it doesnot display any results.
I would like to AUTOMATICALLY count the event for the month BEFORE today
and
count the events remaining in the month (including those for today).
I can count the events remaining in the month manually with this query (today being March 20):
SELECT Count(EventID) AS [Left for Month], FROM RECalendar WHERE (EventTimeBegin >= DATEADD(DAY, 1, (CONVERT(char(10), GETDATE(), 101))) AND EventTimeBegin < DATEADD(DAY, 12, (CONVERT(char(10), GETDATE(), 101))))
Could anyone provide me with the correct syntax to count the events for the current month before today
and
to count the events remaining in the month, including today.
Hi. I'm trying to calculate the total number of cases that more than 4 events. This is my query:
select count(defendanteventpros.vbkey) as countcase from defendanteventpros where exists (select * from defendanteventpros where eventid=2 and eventdate between '1/1/2007' and '12/31/2007') having count(defendanteventpros.vbkey)>4
The defendanteventpros table has a vbkey (primarykey) eventid (type of event) and an evetndate (when the event takes place).
What I'm trying to find is how many cases had 4 or more events and were received (eventid=2 from above) between 1/1/2007 and 12/31/2007.
I keep getting a big number that I know is not correct. I would appreciate any help anyone can offer! Thanks!
We keep a database of events that are classified by severity. I've got a good, efficient query that gives me the grand total for these events. And I have a bad slow query that breaks down the information into daily totals.
My fast query will count the totals for 3213586 events in 4 seconds. The slow daily query takes about 60 seconds for each day.
Please help optimize my slow query!
------ Database schema is:
Column Name Data Type Length Allow Nuls
1 msrepl_tran_version uniqueidentifier 16 0
0 [time] GMTtime (int) 4 0
0 msec int 4 1
0 offset GMToffset (smallint) 2 0
0 type nameType (varchar) 15 1
0 host nameType (varchar) 15 1
0 process nameType (varchar) 15 1
0 dbName nameType (varchar) 15 1
0 ptName nameType (varchar) 15 1
0 description descType (varchar) 47 1
0 rtuName nameType (varchar) 15 1
0 groupName nameType (varchar) 15 1
0 message msgType (varchar) 131 0
0 fgInt tinyint 1 1
0 sevInt tinyint 1 1
0 key1 int 4 1
0 key2 int 4 1
0 spooler tinyint 1 1
------- The database view eventView that's used by my queries: SELECT DATEADD([second], [time] + 60 * offset, '01/01/70') AS timeStr, msec, host, process, dbName, ptName AS point, description, rtuName AS rtu, groupName, message, type, sevInt, [time] FROM dbo.summary
------- Fast totals query:
select (select count(*) from [event].[dbo].[eventView]) as Events, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 1)as Low, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 2)as Medium, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 3) as High, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 4) as Low1, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 6) as Medium1, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 8) as High1, (select count(*) from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 9) as High2
------- Slow daily query:
declare @myDay datetime declare @begDay datetime
declare @myTable table( myDay datetime, Events int, Low int, Medium int, High int, Low1 int, Medium1 int, High1 int, High2 int )
select @myDay = getdate()
select @begDay = convert(datetime, '10/01/2006')
while @begDay <= @myDay begin
insert into @myTable (myDay) values (convert(varchar(20), @begDay, 101))
update @myTable set Events = (select count(*) as Events from [event].[dbo].[eventView] where convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set Low = (select count(*) as Low from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 1 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set Medium = (select count(*) as Medium from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 2 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set High = (select count(*) as High from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 3 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set Low1 = (select count(*) as Low1 from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 4 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set Medium1 = (select count(*) as Medium1 from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 6 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set High1 = (select count(*) as High1 from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 8 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
update @myTable set High2 = (select count(*) as High2 from [event].[dbo].[eventView] where type = 'alarm' and sevInt = 9 and convert(varchar, timeStr, 101) = convert(varchar(20), @begDay, 101)) where myDay = convert(varchar(20), @begDay, 101)
A column in my table is being updated and I would like to find out what process, host, user, etc. is performing the update.
I have added a trigger to the table so that when an update occurs I am able capture the session details into a table, however the session details that are being returned are those of the actual trigger rather than those of the original query.
The statement that I am using to capture the session details is...
select * from sys.dm_exec_connections cross apply sys.dm_exec_sql_text(most_recent_sql_handle) where session_id = @@spid
There is a parent_session_id returned in the above which I hoped might give me some pointers but this is always NULL.
I have been working my way through the system dynamic views in the hope that one of them will provide the information I am looking for but so far I've had no luck.
Because the system I am troubleshooting is web based the actual server and user details will be those configured as service accounts for the application but if I am able to show the developers the original query that was used for the update I'm sure they will be able to provide valuable information as to which component it is coming from. I imagine SQL will also be able to tell me which host is sending the query. With these two bits of info I'm sure I'll be able to track it down.
Our suspicion is that an old component is running somewhere and I'd like to be able to track it down. Perhaps I've spent too much time looking at this but I'm coming to the conclusion that it may not be possible from the SQL back end, which I am very surprised at.
I have tried capturing every session that was running at the time of the update but there's simply too much info to deal with. I even searched through these results but there's so many different possible combinations that an update statement might take form that it quickly became a worthless exercise.
Perhaps one of you can resolve this for me. Any ideas?
I have a sql server express 2005 database with a table named Events with a column named Date (datetime datatype). I want a query that will display all rows that are either current or future but not past. I suspect there is a simple way of doing this. As a Newbie searching archived threads this is what I have come up with so far. I determine the number of days from present:SELECT DATEDIFF(day, Date, GETDATE()) AS NumberOfDays FROM EventsThis yields number of days from present with positive numbers in the past and negative numbers in the future. Thus setting a WHERE clause to <= 0 would limit my results to present or future events. Something like this:SELECT * FROM EventsWhere DATEDIFF(day, Date, GETDATE()) AS NumberOfDays <= 0The error message states: "Incorrect syntax near the keyword 'AS'"This feels like a clumsy way to approach this problem, but I have to start where I am.Any suggestions on how to proceed will be greatly appreciated.
I am investigating using Query Notifications - a great addition BTW. I have built a test app - loosely based on MSDN example - and am running against a SQL Express 2005 server. I have the following problem:
I have start/stop buttons to enable change checking, and a hardcode query that I am using for testing. If I stop and restart I now recieve duplicate notifcations, a single change causing the OnChange event to fires twice with two different ids Repeat this and the event will fires 3 times for each change and so on.
This only happens with a running app, if I restart the app I only get a single notification so I assume I have missed a step in stopping query notifcations or reinitialisation. I have include code below:
private bool Start()
{
try
{
// Remove any existing dependency connection, then create a new one.
SqlDependency.Stop(_currentConnectionString);
if (SqlDependency.Start(_currentConnectionString))
{
if (connection == null)
{
connection = new SqlConnection(_currentConnectionString);
}
if (command == null)
{
command = new SqlCommand(_sqlQueryString, connection);
}
return (true);
}
}
catch (Exception e)
{
MessageBox.Show(e.Message);
}
return false;
}
private void Stop()
{
SqlDependency.Stop(_currentConnectionString);
if (command != null)
{
command.Notification = null;
command = null;
}
if (connection != null)
{
connection.Close();
connection = null;
}
}
private void GetDataSnapshot()
{
// Empty the dataset so that there is only
// one batch of data displayed.
dataToWatch.Clear();
// Make sure the command object does not already have
// a notification object associated with it.
command.Notification = null;
// Create and bind the SqlDependency object
// to the command object.
SqlDependency dependency = new SqlDependency(command);
dependency.OnChange += new OnChangeEventHandler(dependency_OnChange);
using (SqlDataAdapter adapter = new SqlDataAdapter(command))
HiIf I register my own proc via execsp_addextendedproc 'xp_myproc', 'xp_mylib.dll'.How then I can look all my own procs and dll's?In which tables are this data saved?Thx.
