Welcome to another installment of the cloud performance updates. Last quarter we covered improvements in the Single Page Framework and reduced bundle sizes. This quarter, we’ll be focusing on continued improvements to the Single Page Framework for core Jira jobs, as well as a new performance feature that allows Confluence pages to load even faster.
Improvements to Jira SPA
Your dashboard is the main display you see when you log in to Jira. You can even create multiple dashboards for different projects, or multiple dashboards for one massive overview of all the work you’re involved with. Given its prominent use and viewing, we’ve added Jira Dashboards as part of the Single Page Application (SPA) and users are now experiencing in-app transitions at 3.9s at p90, 1.9x faster than the full-page loads they had experienced previously. SPA allows for a more efficient subsequent load, as the browser only needs to process incremental changes for each user interaction. This means that once you open Jira or Confluence for the first time in your browser, navigating and using the apps will be faster.
Improvements to Jira backlog
Backlogs now load 1.4x faster. This includes team-managed and company-managed backlogs (formerly known as NextGen and Classic). The main reason for the faster times is reducing bundle loads, which helped optimize and remove elements of the page that led to longer load times. As part of this improvement, users will notice the backlog will partially load, starting with items such as tickets, allowing users to engage with the page while less-often used features continue to load.
Improvements to Jira roadmaps
We’ve also improved one of our most popular features to date – Jira Roadmaps is 1.5x faster! As a user, you’ll experience this as a 1.3s improvement for page loads. These improvements were achieved by optimizing a hand-picked set of existing components that streamline the work a browser needs to do to display the elements needed on your Jira Roadmaps. Some additional benefits you’ll experience as a result of these optimized components include:
- 2x improvement in changing timeline mode
- 1.4x improvement in expanding/collapsing an epic
- 1.4x improvement in filtering issues
- 1.3x improvement in creating issues
| Action | Load time improvement in the past 6 months |
| Opening a Jira Backlog | 1.4x faster |
| Opening a Jira Board | 1.2x faster |
| Creating a Jira Issue | 1.3x faster |
| Loading a Confluence page (initially) | 1.9x faster |
| Loading the edit button on a Confluence page | 4x faster |
Improvements to Confluence Server-Side Rendering
The Confluence team has also been hard at work for the past several quarters to launch Server-Side Rendering (SSR), a huge foundational piece that delivers our biggest ever performance improvement in Confluence Cloud. SSR speeds up initial page loads (for example, opening a Confluence page in a new tab) so you can start reading and scrolling pages faster. SSR can be especially helpful in scenarios with a slower network or older machines, which may be more common as companies embrace a remote workforce.
When it comes to measuring performance, the metric we hold ourselves accountable to is First Meaningful Paint (FMP) – essentially, how long it takes before users can start reading the page. SSR allows us to make significant improvements in FMP with most page loads now 2x faster than before. Want more details? Check out the full write-up for details on what’s happening behind the scenes to make this come to life.
With this first step in improving the FMP number and our first round of SSR improvements in place, we are now working on adding even more commonly used features to this platform. Our next focus will be on providing similar improvements for images, page trees, and Quick Search. While 1.6-2x faster page loads are good, we know we can do even better. This is the first step in a journey for the Confluence team and we will continue investing in faster and improved performance.
As always, this is just another step toward the continued improvement in our cloud performance. We will continue sharing the work being done. You can always track our progress on the performance cloud roadmap, or join the conversation in the Atlassian Community.