dbi Blog

After 2 years, or even 3 years? The organizers could not agree,

the SQL Conference took place again in Hanau.

73 speakers have taught us in over 80 sessions with their experience as well as knowledge new. There were incredibly many exciting topics. I would like to highlight the sessions “Migration to the Cloud – What could go wrong? and “Performance improvement of Transact-SQL queries without code changes”.

Migration to the Cloud – What could go wrong?

It was exciting to hear how a simple migration to the cloud can become really difficult due to wrong decisions.

A still lively and complex infrastructure, especially used for reports, had to be accompanied into the clouds without AzureDevOps.

How the loading routes were redefined, the security architecture was adopted but also all the structures were tried to be adopted 1-to-1.

All these experiences were shared with us during the session, as well as always in the breaks through conversations.

Performance improvement of Transact-SQL queries without code changes

Let’s imagine, we are sitting on the couch on a Saturday evening, have just baked a nice pizza and are really up for it, to watch a football game.

Suddenly a customer calls us, the SQL server is completely overloaded and he is looking for help.

Ok, beer and pizza on the side, laptop ON.

It quickly becomes apparent what the overload is. The CPU is at 99% utilization. Start analysis.

We find out that a query is causing the high CPU load and have to look for a solution. The query could definitely be better, a proposal change including reindexing was sent to the customer. The answer is, no idea, we have to clarify first. And as we all know, this can take days.

A lot of days. So we need an adhoc solution.

One thing we have already done well, the QueryStore is already running. Now we have possibilities through stored procedures like sp_query_store_set_hints which allow us to influence queries without changing application code or database.

In summary, it was 2 incredibly learning days. I also liked the community, which made a very pleasant exchange possible.

Of course we must not forget Tame, the DataMonster. Thank you very much for the great support. I would love to come again next time.