On November 19th, 2020, I delivered a short session about Azure SQL Analytics, as part of the MVP Days Israel 2020 event.
MVP Days Israel 2020 was a full-day event driven by the local (Israeli) MVP community to share knowledge on various Microsoft products across the board – Azure, GitHub, DevOps, Power Apps, AI, Data Platform and more.
We had a bunch of impressive talks given by very talented people. My session was delivered in Hebrew, and it was mostly based on what I wrote in one of my previous blog posts: Is Azure SQL Analytics all you need for SQL Server Monitoring?.
Due to popular demand, I plan on delivering my session on additional platforms. So, for a start, I’ve set up a new page dedicated to the new webinar, including the slide-deck available for download:
▶ Performance Monitoring with Azure SQL Analytics
Personally, I was a bit disappointed that the talks were relatively short (a measly 20 minutes for the session + 10 minutes for Q&A). It was a far cry from the 40-90 minute sessions that I got used to in our usual Data Platform webinars (in Virtual Groups, SQLSaturdays, Data Platform Meetups, etc.). I believe it was the first time ever that I delivered such a short session!
To top it off, we even had technical difficulties at the start that prevented me from sharing my screen, wasting precious 7 minutes out of the total of 20 minutes (that’s almost half the time!). But with some MacGuyver-level creativity, we finally got it working – with 13 minutes left on the clock!
Naturally, this got me pretty stressed out, and I blazed through my presentation and demo, speaking so fast that I didn’t even stutter (as I usually do). At the end it felt like I just sprinted a marathon. Luckily, the audience didn’t actually feel it in my presentation.
But unfortunately, there were a couple of important talking points that I forgot to focus on. Even more embarrassingly, these were talking points that I touched on at the start, and promised to explain in detail later 🤦♂️. I am so terribly sorry that I didn’t… By the time I reached the end of the demo and Q&A, the moderator even told me that I have some more time and asked if there’s anything else I’d like to add – but I totally forgot about the missing topics that I teased on earlier.
So anyways, I’ll try to fill in those missing blanks here:
Security Auditing – Why do we need it?
In my session I mentioned that it would be very helpful to also add Security Auditing for your database, not just the SQL Analytics Diagnostic stuff. I promised to explain why later, but didn’t.
The reason? It has to do with the fact that the data saved by SQL Analytics is incomplete. It’s missing the actual Query command texts (instead only saving the Query Hash for most events), and missing session-level details such as Client Host Name and Login Name.
The SQL Security Audit has this kind of data, which can be used to “complete” the details missing from SQL Analytics. Correlating the two is not an easy task, and it may become quite tiresome. But it’s still better than nothing.
SQLInsights – What’s that?
SQLInsights is one of the diagnostic types that I promised to focus on later, but didn’t.
It’s actually the most interesting feature in Azure SQL Analytics.
In short, these are intelligent root-cause-analysis and recommendations powered by Artificial Intelligence, that correlates data from different events and metrics to reach an informed conclusion.
Microsoft Docs provides a comprehensive list of these Intelligent Insights here:
Unfortunately, I wasn’t able to generate such example insights for my demo, but I did have the chance to witness this in action in one production environment. It was quite accurate in pointing out issues with the database.
Conclusion
After the session, I updated my slide deck to include the missing talking points that I forgot to explain in detail. It’s uploaded to the dedicated webinar page here.
Hopefully, next time when I’ll be delivering this session, I’d have more time to go into detail and won’t miss anything important.
Once I schedule such a session, I’ll share it in my usual social network channels, so stay tuned!
Pingback: Is Azure SQL Analytics all you need for SQL Server Monitoring? – Eitan Blumin's Blog