VictorOps is now Splunk On-Call! Learn More.

Establishing a Complete Database Monitoring System

Brad Griffith November 04, 2019

Monitoring & Alerting Collaboration DevOps
Establishing a Complete Database Monitoring System Blog Banner Image

Observability and collaboration are commonly referenced as two of the most important aspects of software delivery and IT service management (ITSM). Data flows between services, applications and infrastructure, fueling the systems built and maintained by DevOps and IT teams. Resilient, fast data pipelines and the constant uptime of connected servers and applications are a must in order to remain competitive in today’s modern world of DevOps, CI/CD and complex cloud-based services.

So, how do you actively institute database monitoring solutions and tools that drive visibility into system health, improve data resilience and lead to faster incident resolution? This article will guide you through all the things you should think about with database design and how you can establish a complete database monitoring system for DevOps and IT operations.

The five W’s of database monitoring

The who, what, when, where, why and how of database monitoring is crucial to a successful system. Every engineering organization is built a little differently and accountability for maintaining different services within a system can fall under multiple different teams. So, it’s hard to establish a complete set of best practices for database monitoring. Instead, it’s better to ask the five common “W” questions regarding database monitoring and use the answers to determine best practices for your specific team.

Let’s dive into the key questions you should be asking when researching database monitoring tools and team structures.

Monitoring & Incident Response

Who manages database monitoring?

Well, who’s currently responsible for ensuring data resilience and uptime of crucial servers? Is it a mix of people and teams? Is it the database administrator (DBA)? Should the monitoring responsibility fall to a different team than the team performing day-to-day maintenance and upkeep of data infrastructure? Would you find more value by assigning database monitoring to an SRE team instead of pushing more responsibility onto the data engineering or IT operations team?

Because data persists and flows across multiple areas of a service, database monitoring likely won’t fall under the purview of one team or person. The concept of DevOps is about improved collaboration, transparency and shared accountability for all engineering and IT disciplines – including database monitoring. So, you should really be asking how can you share context across multiple teams to consistently show the health of your data pipelines? With the right amount of context and an understanding of who owns which piece of database monitoring, you can start to build more resilient services and get more value out of your monitoring solutions.

What to look for in a database monitoring tool?

When you’ve established who’s responsible for database uptime and monitoring, you can begin to assess database monitoring tools and processes. Of course, you need to first look at your database to understand the types of monitoring you need. Depending on your infrastructure, you’ll most likely need some combination of SQL server, MySQL, NoSQL, Oracle database monitoring, etc. At a minimum, the database monitoring tool(s) you use will need to do the following:

  • Monitor query details (CPU, disk, memory, etc.)
  • Track replication details
  • Assess overall database performance (buffering, caching, latency, connectivity, etc.)
  • Review scheduled jobs
  • Analyze session details (current user connections and locks)

What are the biggest gaps in knowledge across your data infrastructure? What can you stand to learn from the way users experience your product? Are there database metrics you need to know about in order to make better business decisions? The best way to decide on the right database monitoring system for your team is to identify the pain points you have and implement monitoring solutions to create observable databases and improve the visibility into overall database health.

When do you need database monitoring?

No matter the scale of your business, you need some level of database monitoring. Data resilience and effective data management are crucial to any DevOps or IT organization. And, it’s also crucial that business and legal teams know what’s happening with their data in order to ensure accuracy when talking to customers. But, it’s a constant balance between budget, business need and over-tooling. Database monitoring should always be top of mind but smaller services most likely need simple uptime and high-level performance monitoring.

Where to look when building a database monitoring system

Most homegrown solutions for database monitoring can’t provide you with all the context you need. So, you need a combination of homegrown solutions and purpose-built database monitoring tools to create holistic visibility into service health. If you don’t know where to start, the below tools are a great place to go when you’re assessing database monitoring tools:

Why do you need database monitoring?

The information from your system’s databases fuels the fire keeping your business afloat. The consistent, reliable flow of data between services, applications and servers will ensure that your product behaves in a way that users expect. You’ll provide optimal experiences to customers and be able to provide greater value to end-users. Database monitoring can help you identify problems faster and give you the context you need to remediate incidents in a timely manner. With the right context and appropriate tools for collaboration, you can maintain greater levels of uptime and mobilize incident response teams faster.

How to establish database monitoring best practices

There’s no single system for database monitoring. Best practices in database monitoring simply involves asking the right questions and understanding how to eliminate blind spots. With more visibility into database operations, you’ll know more about the way your system functions and can build monitoring and incident response processes around that knowledge. Observability into database activity helps you build more resilient data pipelines and facilitate more consistent performance across all applications and infrastructure.

See how DevOps organizations are taking action on their database monitoring systems. Sign up for a 14-day free trial of VictorOps or request a free personalized demo to start ingesting monitoring data into a collaborative alerting environment for rapid on-call incident response – driving down MTTA/MTTR and improving end-user experiences.

Let us help you make on-call suck less.

Get Started Now