The On-Call Edition!

Whatever you celebrate this time of year, or even if you don’t, I hope you are enjoying yourself to the fullest! I also hope you remember to write 2024 on all your checks ??

SMEAT (SQL Meat):

How to Make On-Call Better

Assuming you are in IT, you are working every day with stuff that breaks. Hardware breaks, code falls apart at scale, databases get deleted (oops) or corrupted, etc. Very little of this is life and death, but it can certainly be revenue affecting to your employer/client. Revenue is what gets you paid.

This is more of a non-prioritized list than a process, and some of it won’t apply to you.

1 - Understand the on-call SLA.? Know the response time requirements by heart, and commit to being able to meet them.

• This is everything from answering the call, to acknowledging the issue, to working on it. If there is a “resolution SLA”, you may by set up for disaster. Most tech emergency fixes cannot be nailed down to ‘x’ hours
• Also know what qualifies…don’t be shy about telling the caller a ticket is Sev B, rather than Sev A

2 - Get the info about on-call requirements before you start at that job

• How many people in the rotation?
• Response times / SLA?
• Primary and secondary?
• What qualifies for on-call and when can you say no?

3 - Set a different ring-tone for on-call calls/contacts

• On my iPhone I have my Pocket DBA? clients on “Emergency Bypass” and all the authorized callers are in my contacts with full name and company name.

4 - Write KB articles so the person/team/helpdesk that is calling you can at least have some useful info, rather than waking you up with “something bad happened, please fix it”

5 - Above All (yet its at the bottom of the list) – ROOT CAUSE EVERYTHING!

• In my LinkedIn post yesterday (link below) I mentioned dropping ticket counts. We did this by approaching EVERY ticket, on-call or otherwise, from a mindset of “How can we make this call/issue never happen again?”
• Backups failing due to disk space? Get more space or fire the backups directly into a fileshare/cloud bucket
• Jobs all failing at 3AM? Might be as simple as offsetting the start times. Might need some tuning to reduce resource requirements.? Query Store is your friend. Logging sp_whoisactive to a table is your BEST friend!

Do the work. Write the documentation.

A lot of shops measure the level 1 team on “First Touch Resolution” – they get better raises and bonuses by NOT calling you in the first place. Help them help you!

What would you add to the list? Tell me!

(This on-call nonsense happened to me waaaay back in 2007 or so)

My recent LinkedIn post(s):

I've been on call since 2006

My Focus is Horrible!

Interesting stuff I read this week:

The Growing Inequality of Who Gets to Work From Home

A different take about on-call in tech

SQL Server UPSERT Patterns and Antipatterns | Michael J. Swart (michaeljswart.com) – Getting away from MERGE

Free! My YouTube channel offers some barely-edited, one-take videos on SQL Server. Most are 10-20 minutes and cover only one topic.? Many have live demos in them. Most of them are aimed at Accidental DBAs, developers, sysadmins, etc.

SQL tidBITs:

SQL Server FULL backups ignore unused space in the data file. If you have 10GB of data pages in a 100GB .mdf file, you get a 10GB backup, before any compression.

Sponsored by: