cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

butch Job monitoring

suresh230591
Advisor

Hi,

We have a usecase to monitor the batch job which update the certain records in the CRM system. I can see these jobs as web request under one of the service monitored by DT. Operations team need alert on the job failure or if jobs are timeouts. We can see the status and response time of these job in DT.

Now the challenge is that these are not triggered frequently so it is like

  1. once in a day, once a week like wise.
  2. name of the request seen in DT are not static as jobId append at the end so it will always be a new request in DT.
  3. DT mark this as rarely occur request in 72 hour.
  4. We can't mark them as Key request as name always going to change

How we can alert on failure of such request/job?

Please note there is no agent installed on CRM. CRM service is being monitored remotely from calling system where we have full stack agent


Thanks

Suresh


2 REPLIES 2

rodrigo_alvare1
Dynatrace Pro
Dynatrace Pro

Hello,

For What I understand the request will have different names but will be on the same service. And also you will be triggering just one job on that service. If that is the case, I think your best solutions would be to create a custom alert https://www.dynatrace.com/support/help/how-to-use-dynatrace/problem-detection-and-analysis/problem-d...
or modify the configuration from the service to use a static threshold at service level, instead of the baseline deviation https://www.dynatrace.com/support/help/how-to-use-dynatrace/problem-detection-and-analysis/problem-d...

If you still want to get Dynatrace to record the baseline of that service, I recommend you to set some maintenance windows during all the time the batch process is not running.

Regards

JamesKitson
Dynatrace Leader
Dynatrace Leader

Another thing to consider is adding a naming rule to the service so that the name does not appear different every time the job runs. Then you can mark that as a key request and adjust anomaly detection for it.

If there is a 'run id' or something like that you'd be interested in in the 'original' name then you can define a request attribute to pull that out.