On Fri, 2012-03-09 at 11:29 -0500, McKown, John wrote: > notification of events happening on "servers", especially z/OS
We're small like you, lights-out after 9:00pm, and the operations automation we have is all homebrew. I have a console monitor application, written in REXX, that watches for specific messages such as ABENDs. It also watches for messages with MDBGHOLD=YES that remain un-DOMed for longer than a set period. Normal action is to send email via Lionel's XMITIP, and then to escalate to another email address if the first recipient doesn't acknowledge (by simply hitting 'reply' on their email client) within ten minutes. We have another application which runs as a final step of most of our production jobs, and emails the responsible programmer(s) in the event of ABEND or untoward return codes. An MPF exit backs up that application in case the failed job had a JCL error or otherwise could not continue to the last step. A third application analyzes our IDMS transaction log and filters for unusual activity, emailing some of us who care about that sort of thing. Automated email works great, mostly, and can include more diagnostic information than a tweet (say) can include. When email doesn't work - and it isn't foolproof - we rely on our users to call. "Press 1 to wake somebody up." -- David Andrews A. Duda & Sons, Inc. [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
