It's all about the answers!

Ask a question

Practical limits for number of work items in a project area and repository - gathering info

Anthony Kesterton (7.5k7179136) | asked Feb 01 '13, 7:38 a.m.
I am trying to gauge the practical limits for the number of work items that be used in a RTC project area and a RTC repository.  I do realise this is limited by the performance of the underlying hardware/database/app server, etc - but it would be interesting to get a view of the size of systems out there.  I am interested in the RTC version, underlying hardware, database and  app server, and the largest number of work items you have in a project area, and the total number of work items in the repository.

If anyone has seen this info anywhere, a pointer would be most helpful.  If you have a system that is coping well with large numbers of work items, I would be interested in that too.

many thanks

3 answers

permanent link
Sudhakar Frederick (80113531) | answered Feb 14 '13, 6:26 a.m.
Hey Anthony
I had  a system that coped fairly well with about 3 million work items, 300 project areas, 3000 registered users, 50+GB SCM data in 120 components and 300 streams. It was a dual-tier setup on gold instances (8 vCPU, 16GB RAM each) on the Rational Demo Cloud. We didn't get a chance to do any formal performance testing on it (see
(J)Metering a Jazz server – Part I ) simply wishing to see how it coped with such large numbers for query and work item creation.
 It was a temporary setup so not sure if it counts:-)

permanent link
Vivek Pandey (4321236) | answered Feb 15 '13, 1:34 a.m.

hi Freddy,

Was there any visible impact of no# of workitems on OOTB  birt reports ?

permanent link
Sudhakar Frederick (80113531) | answered Feb 17 '13, 5:47 p.m.

The only report we tried was the open vs closed WIs and it wasn't significantly slower. Most of the other "testing" was via JMeter and some simple Web UI gestures, like hitting the Projects page to see how 300 projects would affect it.

Interestingly this setup on the Cloud showed better reponse (when accessed from the local network)  than a similar setup on the local network. Possibly because of excellent connectivity between the db and app server instances on the Cloud vs the local setup having to deal with normal traffic on that link.

Your answer

Register or to post your answer.