MySQL Business-Critical Support — Three (3) Minutes Response Time Challenge!

How on earth does Continuent maintain a less than three (3) minute Average Response Time for Urgent Support Requests?

When was the last time you had an Urgent Support Request related to your MySQL HA/DR/Geo-scale deployment? And did you pay attention to how fast your provider could reply to you?

For most of Continuent customers, the answer to Question#1 is: “It has been for a while.”

We get an average total of 36 Urgent Support Requests per year. This is one Urgent Request every ten (10) days or so.

As you can imagine, this is a minimal number of Urgent Support Requests, especially considering the thousands of Continuent Tungsten cluster nodes for MySQL HA and MySQL DR running in production worldwide.

And whether Continuent customers have requested support or not, the answer to Question#2 during 2022 was an average of less than three minutes - 2 minutes and 54 seconds to be exact.

Also, important is that our reply to an Urgent Request will always come from a highly qualified database engineer with 20+ years DBA and Site Reliability Engineering experience, not just some front-line call center representative.

In many ways, it is unfathomable for Continuent to maintain this very high standard. But year after year, Continuent does.

Continuent is a relatively small entity, an agile and highly efficient team. But noteworthy is that we help to safeguard and to secure more than 20 billion dollars of our customers’ revenue streams.

There are two secrets for this: i) the Tungsten software is very clean and works as expected -- ‘nominal’ as the boys and girls at NASA would say — and ii) we have a highly experienced and dedicated team to support our customers.

We have perfected our MySQL HA/DR/Geo-scale solutions over the years. But even more importantly, each minor and small release of Tungsten Clustering and Tungsten Replicator software goes through extensive testing. For example, our current Tungsten Test Suite includes the following:

• Over 150,000 separate tests using various clustering use cases.
• All MySQL (and MariaDB, Percona Server, etc.) versions.
• Countless operating system versions.
• Other variables.

Prior to each release, these tests run 24 hours on 36 servers, thus resulting in 864 hours of testing per set of code changes.

The testing coverage of Tungsten v7 is about +50% more than the testing of v6. For example, we have extended testing for security features. Tungsten v7 security is indeed improved significantly over V6 and before.

Besides the extension of the Tungsten v7 testing in general, we additionally test 40+ different upgrade paths from older Tungsten versions to v7 with different setups and topologies. And believe me, that’s a lot of testing time included.

This thorough testing ensures that our business-critical Tungsten solution works very well. Tungsten is not 100% bug-free, but it is very stable, with new tests being added every day.

While it helps a lot that the Tungsten software works as expected, at times, our customers still require (day or night, rain or shine) that someone will pick up the case. We do - in less than three (3) minutes. Something which you may not have asked for, but something we trust you appreciate. And something which we are dedicated to delivering.

At any given time, we do not have anyone waiting for the call per se. Instead, we have our whole team available to respond and take the helm as needed.

Shared responsibilities tend not to work. But our approach relies on something other than shared responsibility. It is individual responsibility powered by the team. And that has proven to work very well! In minutes of notice.

Go Team Continuent!