Bracketology: March Madness Lessons in Agile Database Management

Mike Kelly is CTO of Blue Medora and GM of SelectStar.

Every year, one of my colleagues takes personal leave on the first day of March Madness to watch the basketball games on television. She’s a University of Minnesota alumna, and this year the Gophers are projected to be one of the 68 teams competing. All over the country, tournament brackets will be filled out by people who will predict their Final Four teams and go nuts as at least one favorite won’t make it past the first or second round.

I’m no bracketologist, but as a casual basketball fan and serious technologist, I have observed that the best teams don’t just score, they also play great team defense, communicate with each other on and off the court, move the ball around and execute great plays after timeouts. You can see where I’m going with this, don’t you?

If we can draw lessons in agile database management from March Madness, my bracket would include the top ways to increase uptime and performance for business-critical databases. Like the granny shot, traditional database management is decades behind. As companies shift to cloud and adhere to DevOps principles to deliver their applications to market, they need their own Final Four of monitoring best practices to implement agile database management.

Run a Zone Defense

Rather than play man-to-man defense -- deploying separate monitoring tools for each database -- companies should centralize database monitoring and play zone to cover everything from on-premises to cloud infrastructure. Most organizations today use dozens of applications. To bring those applications to production might require more than one instance of a database for dev, test and production. Different applications may run best on different databases – from traditional SQL Server to open-source PostgreSQL and NoSQL or even Hadoop and Cassandra.

By unifying monitoring for multiple in-production databases for virtualized, cloud or distributed scale-out environments, you can save time and money with the ability to manage different databases in a single tool. Each database instance appears in the centralized interface with the same look and feel, as well as the deep-dive metrics that are gleaned from each database. Thus, it’s easier to pull and compare metrics to understand where database solutions may be underperforming. Having insight into these key metrics can help you leverage a zone defense into a high-performing operation with the flexibility, scalability and reliability that you need to win.

Share the Ball

In basketball as in IT, don’t be a ball hog. Give your DBA access to infrastructure operations data, and give your IT admin insight into database health. In addition to managing multiple types of databases, DBAs have to keep an eye on the infrastructure that supports these databases – in data centers and the public cloud. For most, diagnosing infrastructure issues means managing multiple tools or working with other teams to understand if infrastructure details like storage usage, CPU utilization, and bandwidth are slowing database performance. Virtualization-aware relationship mapping saves days of troubleshooting by pinpointing issues faster and more accurately, even if the databases are deployed in different environments. By providing equal visibility to consolidated database and infrastructure data you’re eliminating blind spots. No matter their role on the team, players don’t have to beg for the data to do their jobs, they just get it whenever they need it. Constant insight equals consistent, high performance.

Stay in Your Lane But Switch as Needed

Eliminate your need to be an expert in everything by relying on automated notifications and analytics. DBAs can focus on optimizing queries, without having to be DevOps pros or virtualization experts. As your business grows and leverages more tools and applications, your database team’s attention gets diluted further in an effort to support these applications, bleeding resources away from what’s most important to both your customer experience and the bottom line. This problem can be compounded by the need to cross-train team members on new systems, legacy environments, and the multitude of tools used to maintain and monitor them. In practice, this creates an impossible task for the database team to figure out where exactly a performance issue is rooted.

Instead, you can use a system of auto-generated software recommendations and alerts to flag areas that need attention before they compromise end-user performance, such as incorrect security settings or missing backups. Recommendations can take it one step further to provide insight into how you can prevent alerts from triggering again in the future, as well as modifying parts of your environment to optimize performance.  Consolidating platforms also enables you to manage all alerts in a single location or easily disable alerts for your development environment or other less-critical resources.

Draw Up Creative Plays

Whenever performance issues crop up, the typical response might be finger-pointing at another player. The IT teams that are Final Four-worthy understand that problems can only be solved promptly if infrastructure, database and application admins expose their metrics to all IT members who have a stake in providing excellent customer service and support. Making dashboards available to everyone from DevOps to DBAs allows them to understand and execute the same play.

Like the winning coach’s whiteboard, monitoring dashboards always give the team the best chance to succeed. IT teammates get by-the-minute assessments of how their latest code or virtualization implementation impacts overall database and infrastructure health. Enabling deep-dive analyses helps your team troubleshoot performance problems as soon as they arise. Individual database health scores can be integrated with alerts to stay on top of performance problems. Simplifying the management interface to provide one dashboard for both on-premises and cloud allows the DBA to diagnose infrastructure issues immediately, and connect with the right team or tool to correct them quickly.

While I’m not laying any bets on who’ll make it to the basketball Final Four, I can predict that your cloud database will only perform well if it’s surrounded by the right IT monitoring approach. The probability for success increases exponentially if IT teams -- like basketball players -- avoid playing hero ball and, instead, play team defense, share information equally and keep their eyes open for trouble ahead.

Opinions expressed in the article above do not necessarily reflect the opinions of Data Center Knowledge and Penton.

Industry Perspectives is a content channel at Data Center Knowledge highlighting thought leadership in the data center arena. See our guidelines and submission process for information on participating. View previously published Industry Perspectives in our Knowledge Library.