Expand my Community achievements bar.

Join us for our Coffee Break Sweepstakes on July 16th! Come ask your questions or share your use cases on Creative Briefs for a chance to win a piece of Workfront swag!

Performance - specifically browser freezing due to high CPU and high disk I/O

Avatar

Level 10
Hi, We have found a performance issue which I strongly believe is at the client end. This is not cloud-server or network related....it seems to be platform/browser related. On most small projects (less than approx 30 tasks), performance is normal and quite reasonable. However, on larger projects (e.g. 100 tasks), the browser (chrome or firefox) intermittently and significantly freezes when scrolling up and down, even after the initial list of tasks has finished loading. Once the list of tasks has loaded, and the user starts scrolling down the page, the freezing starts to occur. It freezes and unfreezes several times for 1 to 2 minutes (depending on how many lines there are). Whilst the freezing is occurring, you often hear the computer's CPU fan increase in speed as it is trying to cool the CPU. Looking at task manager, you can see increased CPU usage, steadily increasing RAM usage and high disk IO usage. Sometimes the Windows spinning blue circle appears whilst it is frozen. After the initial period of 1 to 2 minutes, the CPU fan quietens down, RAM stops increasing (although it has increased by 500-600MB during the freezing period) and performance returns to normal whilst you remain viewing that project. The issue recoccurs when you move to a different project. This issue is separate to the initial page load performance, which also takes longer with larger projects (greater than 50 tasks). Whilst users would obviously prefer if the initial page load was quicker, they can sort of understand the initial wait. However once the page loads and they start scrolling, the freezing issue really creates the frustration. While it is freezing it makes it very difficult to edit tasks and do normal work. It seems like there is something in the Workfront code which is causing the browser to work really hard - even after the initial task list has loaded. Note that my Chrome and Firefox browsers are both up-to-date. My PC runs Windows 10, Core i7 processor, 8GB RAM and has a SSD so hardware performance isn't an issue. The freezing issue occurs across other users/devices in our organisation. The larger the project, the worse this issue is. Does anyone else have experience with this issue? I have submitted it to Workfront support who said their developers are investigating but no fix has been identified yet. Regards, David.
21 Replies

Avatar

Level 5
David, I'm not seeing this issue. Did WF support give any indication if this is a wide spread issues? Given that this isn't being reported by other customers, the first place I'd focus is on your views. Perhaps you have a custom column that is causing the CPU to get bogged down. Try changing your view and see if this keeps happening. If this fixes the problem, let the support team know which column is causing the problem. This would help the dev team reproduce and fix this issue for you. Workfront Implementation Consultant, CapabilitySource

Avatar

Level 7
Just as feedback, I went to a project that has 310 tasks and I did not experience any issues in scrolling thru the tasks. This is in Chrome. I just tried it in IE, and didn't really see any issues either.

Avatar

Level 2
Try Standard view with no grouping to speed up viewing. If you are using any specific view, share the name of the view and customization done by you. Vikram Potdar Project manager & Workfront Administrator ConnectiCare, Inc. & Affiliates 175 Scott Swamp Road, Farmington, CT 06032 ConnectiCare.com 860-677-3132 (office) 860-655-9075 (mobile)

Avatar

Level 10
Scott, Greg and Vikram - thank you for your feedback. Workfront Support hasn't yet suggested anything regarding the views being the likely cause. I will have a play around with these and see if I can isolate the issue. It's good to know that it is something that could be easily fixable. Regards, David.

Avatar

Level 10
OK - I have done some more testing and discovered the following: The browser freezing issue continues with existing projects, even if I change the task list view to have only one column (Task Name). There is no grouping or filter applied, nor custom formatting on the one visible column. However, if I create a new project with hundreds of new tasks, the performance is fine...no freezing. Even if I use one of our normal views, there is no freezing issue, so I believe this confirms that the view is not the cause. If I create a new project and copy across all the tasks from one of the existing problem projects, there is no freezing issue, so it doesn't seem to be something related to task data. If I create a complete copy of one of the existing problem projects, there is also no freezing issue, so this indicates that it is probably not related to project data. For some reason, the freezing only occurs with the existing projects. I don't know for sure, but the freezing issue may develop over time. It doesn't really make sense, but it's a big source of frustration for our users. I will update our support ticket and see if this helps Workfront support to work it out. Regards, David.

Avatar

Level 5
David, that's very bizarre. Besides what you've done, I think think of any other workarounds. On Mon, May 9, 2016 at 8:21 PM, David Cornwell < globalforum@communitylists.workfront.com> wrote: > OK - I have done some more testing and discovered the following: > > - The browser freezing issue continues with existing projects, even if > I change the task list view to have only one column (Task Name). There is > no grouping or filter applied, nor custom formatting on the one visible > column. > - However, if I create a new project with hundreds of new tasks, the > performance is fine...no freezing. Even if I use one of our normal views, > there is no freezing issue, so I believe this confirms that the view is not > the cause. > - If I create a new project and copy across all the tasks from one of > the existing problem projects, there is no freezing issue, so it doesn't > seem to be something related to task data. > - If I create a complete copy of one of the existing problem projects, > there is also no freezing issue, so this indicates that it is probably not > related to project data. > - For some reason, the freezing only occurs with the existing > projects. I don't know for sure, but the freezing issue may develop over > time. > > It doesn't really make sense, but it's a big source of frustration for our > users. I will update our support ticket and see if this helps Workfront > support to work it out. > > Regards, David. > > -----End Original Message----- > -- Scott Brady, Integration Consultant CapabilitySource, Inc. Phone: 801-919-7375 email: *scott.brady@capabilitysource.com * home: www.capabilitysource.com -We simplify your work, so you can run your business-

Avatar

Level 10
Yep - bizarre is the word! I just got one of my colleagues to do the exact same process on her PC and the result is the same..... Start with a poorly performing project....make a copy of the project, and then the new one performs fast like it should. Switch back to the original, and the freezing returns, RAM increases etc etc.

Avatar

Level 2
Here are a few items that may impact performance: Browser cache - esp Chrome activity - Number of windows / tabs open activity - any video activity - spam / security filtering System (client) CPU intensive activities Disk I/O RAM Note: any or all the above can impact browser performance System (Workfront) Most users are sharing a few large servers with multi-tenant ...so anyone consuming lots of CPU with large views and/or resetting dates for large projects may be artificially governed to minimize impact on the rest of the stack. (need to verify how governance is managed in the shared environment) Most users are doing very simple things (small projects, simple views, simple reports). Not sure the environment was designed for more advanced users (one or two...maybe, but not in any volume). Based on calls with support, I would assume that Workfront does not have tiered performance servers in place to accommodate larger projects (> 500 tasks). I am working with (IMO) moderate sized projects (~ 1000 tasks) and refresh rates can be as long as 30 seconds to 1 minute (maintenance for these projects takes an enormous amount of time). I would be interested in seeing how many organizations are interested in promoting a tiered service environment where either large projects can be migrated -or- the entire client installation can be migrated to ensure QoS.

Avatar

Level 3
We have had consistent performance problems with Workfront (in Chrome - we don’t bother with IE because performance is too poor), especially when managing large (over 2000 task) projects. Since I work for a financial institution, we have strict firewall and other security mechanisms in place. Workfront claims that the issues are related to our security configuration, but doing in-depth cause analysis requires exposure of our network configuration (to support the analysis) that our network support and security group is not willing to provide. So, we have learned to live with a few idiosyncrasies and simply avoid certain activities when managing large projects. Although we have not seen your specific situation, it may be related to your network architecture. Vern R Phipps EDW Release Manager Zions Bancorporation Salt Lake City UT

Avatar

Level 10

Thanks for the points, Vern. We certainly have issues with larger projects too. However, in this case, the strange issue is that on the same PC, with the same hardware, same Chrome browser, same network, etc the performance and client PC resource utilisation (especially RAM and CPU) is dramatically different between one of our "slow projects" and an identical copy of it. It defies any logic that I can think of, but I can replicate it time and again across multiple PCs and at different times of day.0690z000007ZkgeAAC.png0690z000007ZkgtAAC.png

Avatar

Level 10
Thanks Jeffrey. I actually have rarely felt that the production system was too slow from a server point of view. The Custom Refresh Sandbox is of course much slower but I don't generally have an issue with Production server performance. Our issues seem mainly related to browser/client performance. Looking at your list of items that could affect performance, I don't believe that these are an issue for us (although they are all very valid as other causes). i.e. Using chrome, no other tabs open, no video, no other aps open, plenty of RAM (8GB), Disk I/O (SSD), and performance is the same if AV software is disabled. Regarding the suggestion of a tiered service environment, I'd be interested if this became an option but would likely not see the need to pay more for a higher level at this point.

Avatar

Level 10
Quick update on this. A fix has been made available on the Preview Sandbox and it looks like our browser freezing issue has been resolved. Browser performance (other than usual sandbox slowness) is phenomenally improved now. We just have to wait another week for it to get into Production, but this should greatly improve user satisfaction. Well done to Workfront Dev for the fix and thanks to the Support team in helping to identifiy it and facilitate the resolution. Thanks

Avatar

Level 10
Kudos to the WorkFront Development team. This will be a positive for every user. Good job! Eric

Avatar

Level 2
Anxiously awaiting the implementation -----End Original Message----- This message and any attachments may contain confidential or privileged information and are only for the use of the intended recipient of this message. If you are not the intended recipient, please notify the sender by return email, and delete or destroy this and all copies of this message and all attachments. Any unauthorized disclosure, use, distribution, or reproduction of this message or any attachments is prohibited and may be unlawful.

Avatar

Level 2
David, Is there any mention of this fix in the June release ? I can't find anything out for the timing of this release (which is much needed). thanks Jeffrey

Avatar

Level 10
Hi Jeffrey, I have been told that it will be in the June release. This is occurring on the 29th June (USA time I assume). I have told our users in Australia that they should see the performance improvement on the 30th. What sort of issues are you experiencing? The same browser pausing/freezing? We have had some other issues recently with blank/frozen task lines etc. Apparently this fix is still in QA and no release timing is yet known. There are some other fixes for us in the release though apparently, relating to chart values. Regards, David.

Avatar

Level 2
As we bring aboard new project managers, we are seeing a continual stream of complaints about slow performance. The tool requires a significant amount of RAM & CPU capability (esp. on Windows-based machines). Due to the large in-memory footprint, there is a great deal of paging that is required when working with larger projects (> 250 tasks). The additional paging also creates a drag on network throughput (routers are getting hammered though the backbone is fine). Some of our users are experiencing more wait time than work time so productivity is degrading so user adoption is grinding to a halt.

Avatar

Level 10
This sounds exactly like our issue which, I believe (crossing all fingers and toes) is now fixed and will be available to us tomorrow. e.g. By keeping Chrome's task manager open, I can see that when viewing most of our projects the javascript memory usage starts from around 60MB and builds up over a couple of minutes to around 600MB-1000MB per project. The whole time it is building, the browser freezes. Comparing this to the 'fixed' code in the Preview Sandbox, the memory usage barely increases at all when you open a project and the freezing doesn't occur.

Avatar

Level 10
Does anyone know if the software update has already been released? We have not seen the expected improvement in performance yet on our production system. In fact, our users are complaining about receiving many new error messages this morning.