Help in Automating Batch JCL jobs monitoring

Sai Krishnanand · New User Joined: 05 Jun 2023 Posts: 1 Location: India

Hi!

All kinds of advice and ideas welcomed

I am looking for an automation in Batch JCL jobs monitoring, Normally In jobtrac checkpoint we use commands d=j; d=d,k=law and d=kr to monitor failed jobs, late awaitng jobs and waiting rerun jobs

Kindly provide an idea to automate this monitoring process

Rohit Umarjikar · Posted: Fri Oct 06, 2023 10:20 pm

You can buy a third-party tool to get alert about job failures. like Opsgenie

Willy Jensen · Posted: Sat Oct 07, 2023 2:19 pm

There is nothing built into z/os or JES2. Maybe JES3 has something, but I don't know.
Like every automation package I know of, you must trap the job related messages and act upon them. The trapping can be a WTO exit or a subsystem. As Rohit says, you get the best result by buying a commercial package, of which there are several. They can control your jobflow and send alerts if somthing is not right.

If you want to get your hands dirty and do it yourself, then there are a couple of options.

Simplest method is to use the TSO CONSOLE, see 'Appendix C. Writing REXX Execs to perform MVS operator activities' in the 'TSO/E REXX Reference' manual. Though I personally would not do it, I think to is resource intensive, and too uncertain.

Then there are free WTO exits that you can install and use as you please.
I think that the most versatile one is found in file 708 at CBTTAPE.ORG. You can download it, and extract (TSO RECEIVE) to see the contents, member @FILE708 will be the starting point. I think it can trigger a system REXX, which in turn can do stuff like sending messages and emails when a condition arises. If not then I have seen other solutions that can.
I myself have written a packackage, mostly to see if could be done. It consists of a WTO exit for message trapping plus a REXX based monitor to do all the active stuff. It can control started tasks and batch jobs and even have a small scheduler. It seems to work fine in a lab environment, but I have no idea of how it would scale.

Pedro · Posted: Sun Oct 08, 2023 12:22 am

I do not know anything about "jobtrac checkpoint ", so I did not respond earlier.

But if anything is on the table, perhaps use SDSF Rexx API to check the job queues. You can check the queue and also read in the job output to examine it for details.