We use cloud version of Jira and Confluence on South Korea, but it is too slow to the point where we cannot use!
Is it a network issue? or Attlassian server issue? How can I make it fast?
I've been working with Jira and Confluence at various employers for many years now and it has been almost ubiquitous to expect that it is a painfully slow experience. I find it really difficult to believe, as is suggested above, that the issues are related to individual instances of the application.
On top of that, one of the JS files loaded is JQUERY! Which kinda blows my mind given that it is 2020 and not 2008... Sure lots of Wordpress blogs still use jquery but any serious web application left jquery in the dust years ago in favour of far more modern technologies.
Working with Atlassian products is a constant exercise in frustration and always has been. It is like working in slow motion. If I wasn't forced to use the tools by my employers I would never touch these products again. The ONLY reason tools like Jira are still used is because, as far as I know, there is no real competitor in the space that delivers the same volume of features. The second one pops up, a mass exodus off Jira will begin.
Performance just doesn't seem like it is ever a priority for Atlassian. Hell, I just ran atlassian.com through Google Page Speed insights and it scored an embarrassingly low 14 out of 100 for desktop and an even more despicable 4 out of 100 for mobile.
I really hope that Atlassian is taking a serious look at how archaic their entire platform is because every tool or page I visit with their logo on it is radically slower than any other service that I interact with on a daily basis.
I know Atlassian's motto is "Move Work Forward" but it really should be: "Slow Work Down"
It could be any number of things. The best thing to do first is to review this article:
Please make sure to follow all troubleshooting steps to narrow down your issue.
Failing that, you can generate a HAR file and review it to see what might be causing the slowness:
Let us know if you have any trouble.
I just wanted to add some context that our current metrics show almost all Confluence Cloud requests being delivered to the browser (might take the browser a little more time after that to render the page) in about 3 seconds or less. If you're not getting this experience, follow the articles in Shannon's answer to narrow down the cause and open a case with our support team.
Daniel | Atlassian Support
3 seconds or less? I don't know what standards you have at Atlassian, but now I begin to understand. 3 seconds is clearly too slow. This is the order of magnitude where it's not just a minor performance flaw, it's where productivity is severly affected: https://www.nngroup.com/articles/response-times-3-important-limits/
Could it be optimized so that 90 percent of requests are below 500ms and total uncached page load times to be below 1 second for 90 percent of requests?
There are well know UI/UX standards.
Please do read:
"“…users get impatient and notice that they’re waiting for a slow computer to respond… after about 10 seconds, the average attention span is maxed out."
Please note that end users do not care how fast a request is delivered to the browser. The care about the User Experience as a whole. I wrote in another post that having 3 Jira boards open in any browser will take away from 5 GB to 8 GB of RAM, making the slower hardware unusable.
Also, 3 seconds for a request time is a far cry from what most of AWS services is able deliver (single digit milliseconds, if it's hundreds one should seriously start staring in the general direction of your architect team).
+1 on the opinion that 3 seconds or less to load a site is 600% too long. For comparison, Facebook consistently loads below 1 second. But then again, they actually care about performance (GraphQL, etc.) unlike Atlassian. You need to wake up. Like someone said, if another company emerges with the same set of features, then you're done.
Same goes for me. I'm based in New York and a task/sub-task can take 10-20 seconds to open.
It's made it impossible to use during meetings - we can't use JIRA during sprint planning because of the slowness. Everyone has to individually submit changes so that nobody dies of old age while we all sit and watch single sub-tasks be added to the backlog together.
Dear Atlassian team,
This is most definitely not an attempt to compare a proof of concept to a real enterprise grade system like yours.
What I'm trying to say that the modern front-end SPA app can and should be light-weight.
We stumbled about this proof of concept today:
Ivor agreed I could share with you this, obviously with a disclaimer that his project was not intended to be a complete Jira implementation. However - play around with his app, watch the API calls, browser resource usage, see the UX etc.
Maybe it will help you guys with some ideas how to trouble shoot and improve your UI.
And, yes, I know you're not using React :)
Hopefully it helps!
Thank you for following-up on this thread. We'd really love to get more information about the performance issues you are experiencing. It will help us out if you can obtain the troubleshooting information we require per the two articles I sent earlier:
Please send these details to our Confluence Cloud Support Team so we can get to the bottom of this!
If you are on Jira Server instead of Cloud, I would ask that you make sure you have the relevant troubleshooting information from the following article:
Please also include a copy of your HAR file from above. You can then send that to our support team and they can help you out with this.
Thank you for your understanding!
Could you give us some visibility on your current work & roadmap at Atlassian related to having a usable JiRA Cloud experience ? End user performance is the #1 criterion when you use 5hours+ a PM tool.
It's a pity the cloud experience is so bad, as it's supposed the easy way to explore the tool.
You should definitively consider real use cases with people using your tool from Europe, using decent Intel i5 core 2Ghz/4Cores with "only" 8GB" RAM under firefox, and measure yourself the 8s to 10seconds required to display the active sprint or the backlog page.
I have two projects with 15 engineeers and no more than 3000 tickets... It's not what I called "big projects".
Please, do notreply by requesting a support request. JIRA should help me manage my teams, and I don't want to bother to dig more into the reason of this slowness. It's not my job.
I wish you a good day,
For what it's worth, I ran into the issue where the UI was refreshing very slowly—even populating status lists to update a ticket's status was taking several seconds. Also loading comments and other seemingly quick actions. I went into the developer console and cleared out all localstorage, session storage, cookies, and cache data for our Atlassian Cloud site, then reloaded the page and it's working much better now. YMMV.
I'm so sorry you two are also affected by the slowness in Cloud. We're having a look at instances individually to understand what causes the delay.
Could you please raise a ticket with the Cloud Support Team so we can get a look at your instance and understand what is causing the slowness?
If either of you have trouble doing that, just let me know and I'll take care of that for you.
We ended up migrating from Next-gen projects to Classic projects. I'm not aware of architecture differences between the two project types, but it definitely helped us.
The load time went down from 10-30 (!) seconds to 1-5 seconds. Like I said before, 5 seconds is way too long for a modern web app, but at least it's possible to work with Jira now.
Something to really keep in mind if you decide to migrate, like Daniel Woo mentioned - is the user permission scheme.
We had Atlassian team help us migrate from Next-Gen projects into Classic projects. There are some issues but we can leave with them as a good trade-off for the speed.
The Classic projects in Jira Cloud are as fast as our old Jira Server on-prem. We have response times of 1-4 seconds on average. It's not super-fast but we can at least work with it.
If the performance issue is related to Next-gen board as @Rick_Zakharov stated above, it's also important to keep in mind that Next-gen projects behave differently than Classic projects in terms of permissions, notifications and anything that has the concept of scheme.
In this documentation it states that:
Which takes us to a page that contains a diagram that explains how these permissions work for example.
If you have a business requirement that only allows certain groups of users to have access to particular projects, it might be a good idea to, after creating the classic projects to migrate to, consider creating separate permission schemes to each project, configuring each of them according to each project's requirement.
Summarizing it: keep in mind that Next-gen and Classic work differently in terms of how to configure permissions.
Daniel Woo | Atlassian Cloud Support
Hi everyone, posting here will not change much. Atlassian has public issue tracker where based on votes, they take issues in consideration. As such, I have created a performance improvement ticket:
Please go to that page and put your votes. Thanks 🙏
Thanks everyone for answering last week’s question. The winner of the random drawing from those who commented is: @LarryBrock I’ll contact you separately with your prize details. This wee...
Connect with like-minded Atlassian users at free events near you!Find an event
Connect with like-minded Atlassian users at free events near you!
Unfortunately there are no Community Events near you at the moment.Host an event
You're one step closer to meeting fellow Atlassian users at your local event. Learn more about Community Events