I had a meeting with a big customer of mine to show the value of MySQL Enterprise and Cluster CGE and to share with them the latest and greatest features available in our commercial offerings. During the discussion I said: "Imagine if your MySQL database has a problem at 3pm. Thanks to My Oracle Support and MySQL Enterprise Monitor you can quickly resolve your issues and restore the service".
|MySQL Enterprise Monitor|
You could imagine how happy the customer was about my incredible abilities ;)
However, thanks to MySQL Enterprise Monitor, I'm still alive and I'm here to tell the story.
By using correlated graphs we've been able to filter all the events running at that time and quickly diagnose the problem. For those of you who're curious, they had an issue on their advertising network and, at that particular time, all the banners pointed to the portal bringing the number of active MySQL connections to 8000+ and putting a lot of pressure on the database. In the troubleshooting process we also found some regular spikes in the charts, probably related to batches they weren't aware of and they started an internal discussion to find the cause.
At the end I've learned a few things:
- MySQL Enterprise Monitor is an amazing troubleshooting tool and could also be used to start a conversation between DBAs and DEVs
- Never mention, during a customer visit, that the database might have a problem... or this could really happen
- I've a debt of gratitude with the my colleagues in the MySQL Enterprise team at Oracle. Guys, the next time we meet, drinks are on me ;)