Alert from UC4 to email

[Hima Bindu Khambhammettu]

Hello,

We use UC4 for scheduling SAP jobs and we have 9 schedulers defined in the client.
Due to unforeseen reasons, I see that the schedulers were stopped and we dont really realize until we manually check or receive message from Business.
As an improvement to current setup, we are looking for options to get an alert from UC4 whenever these schedulers are stopped due to any reason to our email.

Thanks and Regards
Hima Bindu K

[Frank Muffke]

Hi

this should help you for the first steps...
https://docs.automic.com/documentation/webhelp/english/AA/12.3/DOCU/12.3/Automic%20Automation%20Guides/help.htm#_Common/GettingStarted/Design/GS_DEV_CALL.htm%3FTocPath%3DGetting%2520Started%7CDesign%7C_____3

cheers, Wolfgang

------------------------------
Support Info:
if you are using one of the latest version of UC4 / AWA / One Automation please get in contact with Support to open a ticket.
Otherwise update/upgrade your system and check if the problem still exists.
------------------------------

[Hima Bindu Khambhammettu]

Hi Wolfgang,

I checked the link you have provided in the post and have tried out something.
It seems we need more details on this.

Below are the list of schedulers which are active at present however when one these goes down we should be able to get mail.
I had tried by making one scheduler down and added to notification script however we havent received the mail or notification as we get for cancelled jobs.

Please suggest.

Thanks and Regards
Hima Bindu K

[Pete Wirfs]

We also wanted automatic notification when any of our JSCH objects are down, for any reason, even if someone shuts one down gracefully and then forgets to start it up again.

Our solution is to run this query periodically and throw an alarm if the wrong number of JSCH objects are currently running:

select 'COUNT:' + convert(varchar,count(*)) as NumberOfRunningSchedules
from eh
where eh_otype = 'JSCH'
and eh_client = &$client#​

------------------------------
Pete
------------------------------

[Reed Byers]

At OSU,we were looking for some way to make a schedule email us if it aborted.

We never found a way.

I'm stealing your code... ;)

------------------------------
Reed Byers
Programmer/Analyst
Oregon State University
------------------------------

[Pete Wirfs]

We had someone try to recycle a JSCH to implement a permanent change the same day, but they were interrupted by a phone call, forgot what they had been doing, and never started it back up.  At the time we had no protection, so it went unnoticed for quite a while.

------------------------------
Pete
------------------------------

[Frank Muffke]

Ahoi,

some time ago I posted a watchdog solution in Automic Community called something like "watchdog" that should do the trick.
You can use it as basis for further development.

As far as I remember its a static VARA that holds the object names that must be running and a EVNT.time that checks periodically if they run.

will try to find it for you, nut after 2 migrations it could be kinda challenge :-)

cheers, Wolfgang

------------------------------
Support Info:
if you are using one of the latest version of UC4 / AWA / One Automation please get in contact with Support to open a ticket.
Otherwise update/upgrade your system and check if the problem still exists.
------------------------------

[Frank Muffke]

ahh here you go...

https://community.broadcom.com/enterprisesoftware/communities/community-home/librarydocuments/viewdocument?DocumentKey=0a7e13c3-0b72-45c8-b34f-0cc0ec932de5&CommunityKey=2e1b01c9-f310-4635-829f-aead2f6587c4&tab=librarydocuments

I am afraid the conversation got lost...

cheers, Wolfagng

------------------------------
Support Info:
if you are using one of the latest version of UC4 / AWA / One Automation please get in contact with Support to open a ticket.
Otherwise update/upgrade your system and check if the problem still exists.
------------------------------

[Christoph Rekers]

Hi there

The watchdog is nice but what happens if a user shuts down the watchdag start process? I would start the watchdog from a different client using the CALL-Api or via REST-Call. That "different client" than can only be accessed by admins to ensure that normal users can't "accidentally" shut down the "watchdog start process"

Cheers
Christoph

------------------------------
----------------------------------------------------------------
Automic AE Consultant and Trainer since 2000
----------------------------------------------------------------
------------------------------

[Frank Muffke]

Hi Christoph,

good point!

cheers, Wolfgang

------------------------------
Support Info:
if you are using one of the latest version of UC4 / AWA / One Automation please get in contact with Support to open a ticket.
Otherwise update/upgrade your system and check if the problem still exists.
------------------------------

[Pete Wirfs]

I like the term 'watchdog'.  We call ours "health-checkers".

I appreciate Christoph's ideas.  Some of our health check processes are also susceptible to failure.  But we run our JSCH checker from a JSCH that is owned/operated by the admin (me).

Every case needs to consider all of the risks; When we were on older versions and our WP services could fail due to memory leaks, we had to address the use-case that the entire AE could fail and no one would know (we are lights out on weekends).  So we deployed a PowerShell solution via windows scheduler to monitor the count of active WP processes.​

------------------------------
Pete
------------------------------

[Carsten Schmitz]

For this reason, we monitor many things in Automic by checking actual work results.

We periodically execute light-weight dummy jobs (OS, SAP etc.), then check whether the AH table has a recent entry for that or whether a file written by the job has an up-to-date time stamp.

We then monitor these things with Nagios, throwing alerts when the system doesn't do any work.

Harsh reality has teached us not to trust any of the script functions in Automic for the purpose of a "watchdog". We had processes being alive and at times also reported alive by Automic, while in fact they did nothing useful.

Caveat is, of course, a very slight increase in load caused by the additional dummy jobs.

Best,
Carsten

[Brian Hudgins]

Hi Peter, 
This is why we always recommend when making changes to an active JSCH, always make the change to the definition and the active monitor, the only difference is you can't specify the calendar in the active monitor, otherwise, if the object that was added/changed is later in the day, it will pick it up. If the object is from earlier, it will pick it up at COD and the new definition will be forced to pick up that change. In this case, more bullets are better than 1.

------------------------------
Founder and Partner
Data Center Automation Consultants DCAC
------------------------------

[Brian Hudgins]

Hi Hima,
My team has tried several different alerting functions including watchdog type iterative jobs, using the monitoring of "EXECUTE_ON_END" and several others but through the different environments and version of AE, we have settled on using the following for objects that are supposed to run around the clock, such as JSCHs.
Using MRT, or Minimum Run Time in the attributes tab allows for a minimum hard coded setting of 2355, that way all but 5 minutes right before midnight are covered, it is does not run the MRT, have it execute an email CALL to you and your admins alerting you. 

That is the easiest method, and it is the least likely to change with versions, as well as it is environment agnostic allowing for it to ignore all the changes to the operating systems and different configurations. K.I.S. always should be considered in Automic during implementation.
KR,

Brian

------------------------------
Founder and Partner
Data Center Automation Consultants DCAC
------------------------------