Jira board - Painfully slow

Mary A.Cares

Hello,

 

I have a Jira board which is currently incredibly slow. I have asked numerous team members and they all have the same issue so it is not an internet problem. The board in question contains around 400 cards. Previously used Trello without any issues.

 

Any tips/advice for speeding up a Jira board? It currently takes 2/3 minutes to open and load a card, add a comment as needed and move it to a different column. It's bordering on unusable at the moment and making what should be a quick job, very slow indeed.

 

Thanks very much,

Mary

Nic Brough -Adaptavist-

Jira boards are trying to do a lot of work when you display them, and they're coded in a way that means they spend a lot of time asking the Jira server for data and formatting the response for us humans.

There is a complicated dance between "server does most of the work" and "browser does most of the work" happening here, and I don't want to bore you to death with an essay on it when I could just try to help.

You have a very simple problem of "too many issues". 

Humans can't work with too many issues, we become overloaded when there's more than 100-ish things on a to-do list and when we see that, we instinctively start to build strategies that reduce it to a size we can grasp.  We don't even know we're doing it, but we instantly start to filter, group or discard things on lists when they take more than 10 seconds to read through.

Jira does not do this, it's a computer, it does what it is told, which is "show the board".  However, Atlassian are striking a balance on boards - they don't want the server to do all the work of thinking about what to show to us, but equally, the amount of data it would have to deliver to a browser to get the board and let the browser do all the work is vast.  So they've got a system that gets the server to do the basics and tell the browser to ask for more data if it needs it,

This translates into three things:

• A board will run multiple queries against the same set of issues.  Queries in Jira are heavily optimised, but, by their nature, quite slow (compared with static text)
• Everything a board does is permission checked
• Querys expand exponentially, so the volume of seed data really matters

So, let's say you have a board that says

Project = XYZ, Card colour based on priority and swimlane by assignee.

The project search is fine, it returns X issues.  You have five priorities - that makes the server run another six queries to separate them out.  Then you've got seven assignees - it's now running a query for project, then thirty five more queries to work out how to group and display them.

Now, for your case, imagine doing that for hundreds or thousands of issues.   It's going to be slow and the load is exponential - effort on 1 issue is x, on 2, it is 2x, on 3, 4x, on 4, 8x, and so-on.  So my tips and tricks are

1. Don't use colours, quick filters or swimlanes unless you can do point 2
2. Stop drawing too many issues for humans to understand into the board.  40 isn't a bad number, 400 is.

Obviously, this only applies to the board view, the backlog does not have these problems as it's a list of everything that might need attention, unfiltered.

 

TLDR: simplify your boards by making them less complex.  Either simplify the filters in use, or reduce the numbers of issues shown down to something your humans can actually read.