I answer your top-voted Azure SQL DB and SQL Server questions from https://pollgab.com/room/brento and lose my temper about what you're doing. Here's what we covered:
00:00 Start
00:40 Mattia: Has bulk logged recovery model ever got you across the finish line? If so, could you share a story?
01:41 ForEachDBA: My manager wants to reduce cpu from db servers as cpu usage under 50% most of the time, also sp_BlitzFirst doesn't show cpu top waits. The question:how to measure effectively the impact of reducing cpu? What are the key metrics to watch? Plan cache durations stats is good approach?
02:56 Ataollah S: Will we see any new features added to SQL 2022 in 2024?
03:31 Henrique Almeida: As a recommendation, MS suggests that we use the maximum for MAXDOP being 8.We changed it to 2 and I could see that there was a considerable decrease. give your opinion on the relationship between MAXDOP x CXPACKET, whether they are proportional or inversely proportional.
04:42 Bonnie: Is it ok / safe to simultaneously run sp_blitzcache and sp_blitzindex on a busy boxed SQL Server? sp_blitzindex sometimes takes much longer to complete than sp_blitzcache.
05:45 Owl Kitty: What is the danger zone for Batch Requests per Second that we should be on the lookout for?
06:16 TimelyDBA: My friend is configuring alerts based on severity in a shop requiring the server instance hardened using the center of internet security standards which recommends never using SQL server for mail services. Have you alternative pathways to harmonize security and delivery for such?
07:02 Nazanin B: What are the common issues you see when applications connect to SQL Server over a WAN connection vs a LAN connection? Does compression help here?
07:38 MyTeaGotCold: How can I quality assure SQL code? Other languages have reasonable error handling, good linters, and unit testing that doesn't need to be hacked in.
08:49 Bruno: What is your opinion of the new Azure Boost feature to improve VM network and storage performance? Game changer for SQL VM?
09:04 Bahareh A.: Is column store index maintenance for large tables hard on RPO when log shipping is involved?
09:47 Rudy: SQL assessment API is complaining about high page splits in our canned SQL instance? How do we identity which tables are experiencing this? Is this a fools errand?
10:37 TimelyDBA: Who is the brent ozar for SQL server always on AG and other HA/DR strategies?
12:55 Jen: What are the best / worse use cases you have seen for sp_invoke_external_rest_endpoint to call external API's inside of SQL Server?
13:56 Bonnie: sp_BlitzFirst is reporting a priority 40 finding for Forwarded Fetches/Sec High: TempDB Object for boxed SQL. When is this bad / concerning?
14:14 UnInstallAllScripts: How install scripts like InstallAllScripts.sql gets update?
15:18 Manimal: Does it ever make since to run boxed SQL Enterprise on a 24 core or less cloud VM? Do you see this much in the wild?
15:48 Persephone: Which vendors do you foresee doing cool AI integrations with SQL Server? Which integrations are you looking forward to?
17:14 Rom: What are the top customer issues you see when resizing an Azure SQL VM up or down?
17:49 Dom: Hi Brent, Have you ever taugh of offering your training/blog in french (for Canada/Europe people who doesn't understand english that much)
18:13 Pytzamarama: Happy 2024! Brent I will rephrase a badly asked past question of mine: When we update our customers we drop create all procedures and triggers. Does this affect performance and how. Thank you!
19:50 Zappy: Hi Brent, as far as know, a statistic update shouldn't cause a plan recompilation if a table wasn't modified. But recently I think that I have seen a plan recompilation on an unmodified table simply caused by a stat update with a different SamplingPercent. Is this possible ?
Watch video Office Hours: The Question Isn't The Problem Edition online without registration, duration hours minute second in high quality. This video was added by user Brent Ozar Unlimited 16 January 2024, don't forget to share it with your friends and acquaintances, it has been viewed on our site 2,915 once and liked it 65 people.