BUG: 6.0.1.5108 - Multi-Select DB, ONE gets overridden

PDinCA

For multiple servers targeting 2008R2 up to SQL 2014, the Database Unavailable Alert is to be overridden from "no emails" to just system and production User databases, not those used for restore or side-by-side testing. (This is so pagerduty alerts for inconsequential databases do not fire in the small hours of the night!)

If I select several databases and configure an override, only ONE of the databases shows the override is registered.

I refuse to click over 100 databases across 21 servers.

This is a horrid bug. Please fix ASAP.

UPDATE:

The same is true for "Backup overdue" AND "Integrity check overdue" overrides - select MANY, get ONE. Totally unacceptable for a "relentlessly tested" product.

UPDATE 2:

If I expand, say, the Job failing alert tree so I can see each job by name, then override the configuration, when the page returns, the tree is collapsed! Why? Leave it expanded...... please! This makes EVERY expandable node collapse, making the UI horrible to use.

Dan B

Hi PDinCA,

Thanks for posting on the SQL Monitor forums.

The intended behaviour is to only go up and down the tree in a Server/Instance/Database way. I believe you're trying to change alerts on databases in multiple instances?

Instance1Database1
Instance1Database2
Instance2Database1
Instance2Database2

Is this correct? When I tested locally I could change a select number of databases for one instance but not across instances which is why I have jumped to this assumption. If you are trying to do this on one instance can you confirm which version of SQL Monitor you are using?

If you are going across instances I will discuss with the developers what the intended behaviour in this instance should be.

This should also work for update 1.

I will also raise a bug report for update2.

PDinCA

Nothing at all across instances, Dan.

There is but one instance per server, and they are on independent Rackspace devices, each with their own local domain.

I expand the Alert Configuration object tree to the server and then click the SQL instance, typically (local). That gives me access to the job failing alert's configuration. When I expand that, I see each and every SQL Agent Job, which is the desired result. There is a checkbox alongside every Job, which is also desirable.

For:

1. SQL Backups, I want to send them to a DBA email address. 2. For COM+ Component failure detection, that goes to a very specialized

'complusfailure@us.com' email address.
3. For business process failures, as this is an ERP solution that has a nightly batch process flow, those job failures go to the Support Team.

If I click a set of Jobs, change the alert email, having overridden the default behavior, my expectation given the obvious UI paradigm presented to me "mutli-select+action affects ALL selected objects", you can perhaps see my consternation when I must now go to every one of the 20+ servers, click every job that isn't destined for the default delivery, which means varying quantities of jobs depending on the database count, modules deployed, backup solution in place, and a slew of other variables, just to override the target email. This should be naturally accomplished by the tool.

It gets worse when you add in the overrides for non-production databases, in terms of "silence please" when integrity checks, DB backups and Log Backups yell "overdue". Thus 20+ servers, * 3 alert sets, * N databases to vary from the norm, which, due to the way the tool is designed means that I must set "NO Alerts" as a default, then add-back the system and user databases of interest, just so I can "silence" the ad-hoc restored databases that go offline in the wee small hours, waking me unnecessarily...

Does this paint a better picture of the normal hardships tool-wise that I'm seeking to overcome by use of a multi-select+override paradigm? I somehow doubt I'm alone in this predicament.

Dan B

Hi PDinCA,

Thanks for your patience with this issue.

A support request has been made for this with internal reference SRP-10872. I will update this post when I get some further feedback on this.

Dan B

Hi PDinCA,

Thanks for your patience with this issue.

The issue with multi-selecting has now been resolved as of SQL Monitor 6.0.10. You can now multi select jobs under the job failing banner to disable or modify these.

You can download 6.0.10 from here

ftp://support.red-gate.com/patches/SqlM ... 31Oct2016/