Getting started with Data Center monitoring

This article features some basic guidelines on monitoring the health of your Data Center instance. There are two types of metrics you need to monitor:

Usage

These metrics reflect how your organization uses the Data Center product – number of users, amount of pages (in Confluence), number of issues (in Jira), etc. They also cover cases where your customizations can be quantified – for example, how many custom fields your Jira instance uses.

Infrastructure

Every action your users perform on the product adds load, and the effect of that load can be compounded by how you've customized and configured your Data Center product. Infrastructure metrics reflect how your load affects the environment on which your Data Center product is hosted.

In large organizations, the task of managing the Data Center application, its dependent components, and underlying infrastructure might be distributed across different roles. Understanding the difference between usage and infrastructure metrics can help you identify where each metric is tracked, along with who should be responsible for tracking and analyzing them. 

Usage metrics

Usage metrics can be tracked through a product's administrative user interface or database. This section provides some basic guidance for Jira and Confluence.

Track trends in usage and watch how they grow over time. Use these growth trends to help inform your predictions on how your load will look down the line. The lists below provide the fundamental metrics you should be tracking for both Jira and Confluence:

See Tools for monitoring your Data Center application for related information.

Infrastructure metrics

Infrastructure metrics generally require third-party tools to track, particularly if you're tracking the performance of multiple nodes in a cluster (as opposed to just a single host). Refer to Getting started with Data Center monitoring for examples of tools you can use.

The following infrastructure metrics are fundamental to any monitoring strategy. Track them regularly for each server in your deployment:

Application server

• Average CPU load
  • Per minute
  • Per 5 minutes
• RAM usage
• Disk usage percentage (for all mounts)
• Apache requests  per second (5-minute average)

Database server

The most important database metric to track is the number of open connections (5-minute average). A high number (like 50) can indicate that there is a bottleneck, and may need investigating.

You will also need to track the total number of the following database metrics at regular intervals (for example, every 4 hours):

• Inserts
• Updates
• Locks
• Commits
• Deletes
• Buffer hits

Assuming your database server is hosted on a dedicated server (as recommended), track the following as well:

• Average CPU load
  • Per minute
  • Per 5 minutes
• RAM usage
• Disk usage percentage

Java Virtual Machine

As Atlassian applications are Java-based, you can use many of the monitoring points built into the Java Virtual Machine for insight into the health of your instance. The following metrics are of particular interest:

• Number of available processors
• Heap space
• Garbage Collector activity: see Garbage Collection (GC) Tuning Guide (Jira) for related information.

You can also use JMX (Java Monitoring eXtensions) to monitor your application in real-time. JMX uses objects called MBeans (Managed Beans) to expose data and resources from your application. The following articles provide more details on using JMX for specific products:

• Live Monitoring using the JMX interface (Jira)
• Live Monitoring using the JMX interface (Confluence)

• Enabling JMX counters for performance monitoring (Bitbucket)

Trending and alerting

Once you collect enough data about your organization's usage and infrastructure load, seek out patterns that can help you identify:

• Peak/non-peak hours of usage
• Growth trends

Whenever possible, set alerts for metrics that need further investigation. For example:

• the number of open database connections exceeds 50 over a 5-minute period
• JVM heap space (as in, the amount of RAM used by Java) exceeds 80% of the available capacity

Use an alerting system that administrators and stakeholders can access easily. Your alerting system should also be compatible with your monitoring tool; for example, see Add apps (Stride) for information on how to integrate applications into a Stride room or direct message. 

We're here to help

This article features advice typically provided by Atlassian Advisory Services to customers on forming a performance monitoring plan. Advisory Services can provide strategic guidance to help you choose the right tools and methods to track the health of your instance.