A rather interesting discussion is taking place on LinkedIn about workload automation and mobile devices. We got into a debate about whether the traditional model of having someone watch the scheduling environment 24×7 is still a good use of resources in the workplace. I argue it is not, and that modern workload automation tools (such as BMC Software’s Control-M) can help you streamline your workload automation and reduce the need for people to be watching anything 24×7.
You can read the entire thread over on the Enterprise Job Scheduling & Workload Automation group on LinkedIn. Below I’ve included one of my posts where I talk about my thoughts on the subject matter.
A few things…
[…]
"Watching the jobs": First, nobody at our company watches the jobs. The Enterprise Manager (we use Control-M) is never running 24×7 on anyone’s desktop. I tell people from day 1 that the Enterprise Manager is strictly a "novelty tool to give you pretty pictures". All jobs should be setup to handle known errors correctly, and to escalate and use other tools to notify for unknown errors. We are strictly lights out. Right now we are on Christmas holiday and nobody is "watching" anything in our scheduling environment. It’s all handling itself and if something should go wrong then the tools will take over to notify us and help us handle it appropriately.
I’m not a manager – I’m down in the trenches with the rest of you. However, there is a reason these tools are no longer called "Batch Scheduling" and are now called "Workload Automation". The idea of anyone watching anything is so 1979 it isn’t funny. If you aren’t automating failures and error recovery then you aren’t using the tool correctly. Three years ago I set out on an effort to massively overhaul our environment to automate everything that required manual intervention. Today, we all go home at 5pm and don’t worry about it.
As others have pointed out, no you don’t need the full GUI on your mobile device. However, you need to be able to interact with your jobs no matter where you are when things go wrong. I have absolutely no need to see my entire environment — but when a job errors and automation cannot recover from it I need a menu of options presented to me. For example, today I get an Alarmpoint notification and I can (from my phone):
* Force the JOB OK
* Bypass the job
* Rerun the job
* View SYSOUT (sends it to me via email)I can do this no matter where I am on whatever device I am using. Those 4 commands let me do most of the diagnosis I need to do. It doesn’t solve everything, but it’s been at least 8 months now since I’ve had to wake up in the middle of the night and take care of anything. Compare and contrast that to three years ago when I was lucky to make it through a night without being woke up.
It’s all about how you use the tools – and companies nowadays are looking to take a lot of this stuff to the next level and are investing money to gain efficiencies. If you still have a staff that "watches" 24/7 that is not a job I would want to be in because I can almost guarantee you that those jobs won’t be around in 5 years time.
[…]
