Comments
5 comments
-
Hi Stephen,
You can try plugging this query into a custom metric against the MSDB database. It will report the number of seconds a job named MyJob has been running. You should then be able to set an alert when the number of seconds has been exceeded. I hope this helps!/* Get the JobID for MyJob */ DECLARE @JobId UNIQUEIDENTIFIER SELECT @JobId= job_id FROM msdb..sysjobs WHERE name='MyJob' /* Define a temp table to hold xp_sqlagent_enum_jobs This xp will tell us which jobs are running */ DECLARE @ExecutionStatus table ( JobID uniqueidentifier primary key, -- Job ID LastRunDate int, LastRunTime int, -- Last run date and time NextRunDate int, NextRunTime int, -- Next run date and time NextRunScheduleID int, -- an internal schedule id RequestedToRun int, RequestSource int, RequestSourceID varchar(128), running int, -- 0 or 1, 1 means the job is executing CurrentStep int, -- which step is running CurrentRetryAttempt int, -- retry attempt JobState int ----0 = Not idle or suspended, 1 = Executing, 2 = Waiting For Thread, 3 = Between Retries, 4 = Idle, 5 = Suspended, [6 = WaitingForStepToFinish], 7 = PerformingCompletionActions ) INSERT @ExecutionStatus EXEC xp_sqlagent_enum_jobs 1, '', @JobId /* Get the number of seconds MyJob has been running, if it is running */ SELECT DATEDIFF(ss,a.start_execution_date,GETDATE()) AS RunningSeconds FROM msdb..sysjobactivity a INNER JOIN @ExecutionStatus s ON a.job_id=s.JobId WHERE s.running=1
-
Helps IMMENSELY.
Many thanks, Brian.
UPDATE
Sadly, the join criterion on just job_id gets me the list of every execution, so I revised the query, using my own replacement for DATEDIFF() for the hours calc as I want to be precise, not just have it fire after 3hrs 1min if it starts at, say, 12:59... (Hate DATEDIFF!)
Using SELECT TOP 1 ... ORDER BY a.start_execution_date DESC gets me the currently executing Job details./* Get the JobID for MyJob */ DECLARE @JobId uniqueidentifier SELECT @JobId = job_id FROM msdb..sysjobs WHERE name = '_Quarter Hourly Update Aggregation' /* Define a temp table to hold xp_sqlagent_enum_jobs This xp will tell us which jobs are running */ DECLARE @ExecutionStatus TABLE ( JobID uniqueidentifier PRIMARY KEY -- Job ID , LastRunDate int , LastRunTime int -- Last run date and time , NextRunDate int , NextRunTime int -- Next run date and time , NextRunScheduleID int -- an internal schedule id , RequestedToRun int , RequestSource int , RequestSourceID varchar(128) , running int -- 0 or 1, 1 means the job is executing , CurrentStep int -- which step is running , CurrentRetryAttempt int -- retry attempt , JobState int ----0 = Not idle or suspended, 1 = Executing, 2 = Waiting For Thread, 3 = Between Retries, 4 = Idle, 5 = Suspended, [6 = WaitingForStepToFinish], 7 = PerformingCompletionActions ) INSERT @ExecutionStatus EXEC xp_sqlagent_enum_jobs 1 , '' , @JobId /* Get the number of hours the has been running, if it is running */ SELECT TOP 1 DBA.dbo.udf_GetElapsedHours(a.start_execution_date, GETDATE()) AS RunningHours FROM msdb..sysjobactivity a INNER JOIN @ExecutionStatus s ON s.JobId = a.job_id WHERE s.running = 1 ORDER BY a.start_execution_date DESC
Thanks, though, for your original query - wouldn't be anywhere without it. 8) -
As there's no Uservoice (that I can see), could I register this post as an Enhancement Request?
As it is, I can use the above for ONE Job per Custom Alert, which is fine if that's all one cared about, up to a handful perhaps would be OK, but in environments with many servers and many, many Jobs, I can see this kind of thing being unwieldy in a hurry.
Would RG please consider adding a standard alert "Job Still Running"? It should be reusable by being able to select the job of interest and state the job duration that when exceeded would raise the alert. Alert thresholds could use the standard Low/Medium/High measures for "Duration in excess of the Alert Duration", perhaps.
Thanks in hopes. -
Hi Stephen,
I'm glad I could help. I'm surprised that you can't just get this information from sysjobhistory, but that table only updates after the job finishes. So you have to use an undocumented xp to get the running jobs and then check the job activity table.
Sql Monitor has a UserVoice site here in case you want to add suggestions:
http://sqlmonitor.uservoice.com/forums/ ... uggestions -
Thanks for the Uservoice URL, Brian, I added to the very similar suggestion titled, "specify a time in which jobs should be finished by" at http://sqlmonitor.uservoice.com/forums/91743-suggestions/suggestions/2708577-specify-a-time-in-which-jobs-should-be-finished-by
Would you please post the RSS URL for the SSC Uservoice as I like to keep an eye on suggestions that way (I already have the SQL Prompt, SSC and Search RSS feeds active). Thanks.
Add comment
Please sign in to leave a comment.
I need a Custom Metric that tells me when an active Job has been running for more than my (threshold of pain) max "exceptional duration".
For now, I need that threshold to be in hours. If I can make it data-driven by Jobname, that would be ideal! If the query can poll the same target server as the Job is executing on and retrieve a configuration value from one of my existing configuration tables, I can then set up my "Job Still Running" Alert:
When Job <name> has been running for more than <configured hours for Job <name>> Alert me with High Importance.
Can this be done, and if so, could Red Gate share a sample query for "retrieve currently executing Jobs" that I can filter?
This could then be added to the list on the http://sqlmonitormetrics.red-gate.com site for others to use.
In hopes, with thanks, Stephen