How the checks work v4.6

Most of my servers have sqlwatch v3. I have not updated the ddl since then.
I installed the new version 4.6 on a test server today to see what is new and different.

I see that a lot of the checks logic changed to be specific per database and per job. I see that it created a new check for each individual job to check if that job failed or not.
as a result, I went from having about 25 checks to now having 1,085 checks running!

I’m curious on the thought process behind this. Personally, I like the old way better, sqlwatch would be running 1 query against the database engine to see if any jobs failed, opposed to running 663 queries at the same time to check each job one by one.

I tested creating a new job, and it made a new check for that job. but then after I deleted that job, the check remained. so sql watch will continue querying for failures for that job that got deleted unless I manually clean them up.

I guess with this approach I could disable individual checks if there were specific jobs that I didn’t care if they failed? Is there any other benefit that I haven’t thought of? This seems like it would create more overhead needing to check each job individually, and backups for each database one by one, etc.


There is now a concept of check templates that can be “expanded” by a given object. For example, we can have a check template that checks each agent job, each database etc., so those objects appear in the notifications (“Job XYZ failed” instead of “Job(s) have failed”). Another benefit is more granularity (we can have different thresholds for each).
However, you can continue with the old approach, just ignore the templates, or NULL out the “expand by” column

The checks will also be deleted automatically when the parent object has been deleted, after a certain threshold, as defined in the config table

Granted though, the docs needs updating as its not obvious how it works

Aaahah, that makes more sense now. Thank you for explaining it a bit more. I’ll have to look at the templates more this week.
I needed to tweak some of the queries or views before for things like not warning about a swap drive that is 99% full, or ignoring backups for some replicated databases I don’t back up daily, etc.

Let me know if you need any help. I will be happy to jump on a zoom call with you and answer any questions or help you get going, to make things easier for you.