Jira Service Desk 4.5.x Long Term Support release performance report

This page compares the performance of Jira Service Desk 3.16.6 and Jira Service Desk 4.5 Long Term Support release.

About Long Term Support releases

We recommend upgrading Jira Service Desk regularly, but if your organisation's process means you only upgrade about once a year, upgrading to a Long Term Support release is a good option. It provides continued access to critical security, stability, data integrity and performance issues until this version reaches end of life.

Performance

Jira Service Desk 4.5 was not focused solely on performance, but we aim to provide the same, if not better, performance with each release. In this section, we’ll compare Jira Service Desk 3.16.6 to Jira Service Desk 4.5, for both Server and Data Center. We ran the same extensive test scenario for both versions.

The following graphs present the difference in mean response times of individual actions performed in Jira Service Desk, separated into 3 categories: Heavier actions (takes longer to run), Medium actions and lighter actions (faster to run). Under the graphs, you can click on the "Click to see the data in a table" link to see the mean response times.

The performance was measured under a user load we estimate to be peak traffic, on a 5000 users instance.

To check the details of these actions and the Jira instance they were performed in, see Testing methodology.

In summary

We have performance improvements in almost all scenarios across the product, under high load. This is our fastest version ever! Highlights:

• Viewing queues with SLAs is now 10x faster
• Adding a comment in the customer portal is now 3x faster 
• Viewing queue: all open issues is 3x faster
• Creating a customer request is 2.5x faster 
• Viewing customers/organizations page is now 2.5x faster
• Viewing a service desk issue in the agent view is 2x faster

Overall, we have cut down the response time by 20% to 90% in all scenarios with the exception of "Invite team" with no improvements, and a slight 50ms degradation in "View a customer request on the customer portal". 

We'll continue to invest in improving future performance, so that service desk teams can move with ease through their workspace, and our largest customers can scale confidently.

Testing methodology

The following sections detail the testing environment, including hardware specification, and methodology we used in our performance tests.

How we tested

Before we started the test, we needed to determine what size and shape of dataset represents a typical large Jira Service Desk instance. To achieve that, we used our Analytics data to form a picture of our customers' environments and what difficulties they face when scaling Jira Service Desk in a large organization.

The following table presents the rounded values of the 99th of each data dimension. We used these values to generate a sample dataset with random test data.

Baseline data set

Actions performed

We chose a mix of actions that would represent a sample of the most common user actions. An action in this context is a complete user operation, like opening an issue in the browser window. The following table details the actions that we included in the script, for our testing persona, indicating how many times each action is repeated during a single test run.

Test environment for user actions

The performance tests were all run on a set of AWS EC2 instances. For each test, the entire environment was reset and rebuilt, and then each test started with some idle cycles to warm up instance caches. Below, you can check the details of the environments used for Jira Service Desk Server and Data Center, as well as the specifications of the EC2 instances.

To run the tests, we used 21 scripted browsers and measured the time taken to perform the actions. Each browser was scripted to perform a random action from a predefined list of actions and immediately move on to the next action (ie. zero think time). Please note that it resulted in each browser performing substantially more tasks than would be possible by a real user, and you should not equate the number of browsers to represent the number of real-world concurrent users.

Each test was run for 40 minutes, after which statistics were collected.

Here are the details of our test environment:

Test environment for indexing measures