Question & Answer
How Can you increase performance when loading a large plan using IBM Rational Team Concert (RTC) Web User Interface (UI)?
Accessing large plans in the web UI can cause performance degradation, and it may take a long time to load the plan.
The following items should be considered to ensure the best performance when loading plans within the web UI of RTC:
- In general, plan loading with Mozilla Firefox as an Internet browser has shown to have faster response times when compared to Microsoft Internet Explorer.
- Try to control the plan size to be within 300 - 500 top level plan items (work items) if possible to get the best plan loading performance
- Try to control the total number of work items to be under 500 within one plan to get the best performance
- Reduce the number of displayed columns in the plan. Follow these steps:
- Click on the icon that shows the pencil symbol.
- Select the Edit "ViewName" option, as shown below:
- Select Column Display.
- Select the columns you do not need to display.
- Click on the "x" at the far right to remove them.
- Click on OK to save the View.
1. Click on the icon that shows the pencil symbol.
2. Select the Edit "ViewName" option.
3. Select Options.
4. Select the Group/Sorting options you do not need.
5. Click on the "x" at the far right to remove them.
6. Click on OK to save the View.
- For larger plans, de-select the Include All Items option. If selected, this option Includes execution items from sub-teams and past iterations. For a detailed explanation of what this option does, see details from here
- In Rational Team Concert 220.127.116.11 improvements have been made to both loading work items in a batch, as well as compressing the data to improve performance while loading in the web UI.
- In Rational Team Concert 4.0.1 your plans will render up to 25% faster in the web UI, when compared to Rational Team Concert 4.0. In Rational Team Concert 4.0.3, plan loading performance improved by additional 40 percent. See details from here.
- Whenever a plan is opened or refreshed, before loading work items, the project area configuration data (metadata) needs to be loaded first. Metadata includes full value set of some attributes such as target, owner and categories. Since some attributes have dependencies of each other, like owner is depend on category (file against), and target, category can be mapped to large number of team areas etc, there are large amount of plan loading time is spend on calculating their relationship to pull up the value list. The more categories, team areas, users and target time lines a project area configured to have, the longer time it will take. That says, if metadata is large, it will increase a plan's loading time even regardless the plan size.
- Starting from 4.0.3, users can tune some query string parameters to adjust how the application loads large plans, for example, client can tune the pool size of parallel requests for retrieving plan items, the number of plan items per request, the pool size of parallel requests for retrieving plan item links, the number of plan item links per request etc. Details can be found in here.
- If a plan is configured with the display tree option, the more levels of parent and child display level, the longer the plan to load. In 403, client can tune a large plan by reduce the tree display level for view that has parent - child attribute. By default RTC will load all level of display tree, but client can reduce the display level to speed up the loading process
- Try to disable loading plan details by default.
- Base on our case study, in a 4.0.3 single install CLM server, in a project area with 5 timeline, 5 iteration, 1 team area, 50 item owners, 3 level depth of WI type, loading a plan with 600 WIs without plan details, with a few standard display columns takes 12 seconds in FF, 1200 WIs takes 16 seconds. This can be used as a base line to measure plan loading performance in RTC. See more details about this case study from here
Additional Articles and Resources
Some Planning Performance and Scalability testing data can be found from Jazz.net Task 202728 as reference benchmark data.
|Jazz and Rational Team Concert have an active community that can provide you with additional resources. Browse and contribute to the User forums, contribute to the Team Blog and review the Team wiki.|
Refer to technote 1319600 for details and links.
Rational Team Concert
16 June 2018