Hi Jason,

that is a really tough question to answer. In fact, without knowing the specifics of your data it is almost impossible to answer it in a satisfying way.

I am aware of an OpenProject Application having 150k issues. Issues, because it is still running on OpenProject 2.4 (Ruby 1.8.7.). This installation is lacking performance in some scenarios. For example, a non admin user having about 200 project memberships is opening a wiki page that has a large amount of references (#, ##) to issues of different projects on it. There are a couple of plugins installed for that installation that decrease performance of authorisation checks. In most use cases however, the performance is still ok.

We expect a significance performance gain when this installation is switched to OpenProject 3.0. as we addressed certain performance problems and are switching to a new application stack (rails, ruby) on top of that.

But whether the characteristics of the data in that installation matches the characteristics of the data you are working with I cannot say. So instead of answering your question I have to counter with a whole bunch of question myself:

• In which use cases is the application running slower than before?
  • Which page gets called? What data is displayed on the page (one package, a whole package hierarchy)?
  • What kind of user is calling the page (permissions, roles, project memberships)?
• How many packages are there per project?
• How are the packages related (related to, follows, blocked by, …) to each other?
• What kind of tree structure (hierarchy) do the packages form?
  • Single node tree?
  • Maximum number of nodes in a tree?
  • Depth of the tree?
  • Width of the tree?

Please don’t get me wrong. I am not trying to block of your question by asking questions myself. We are very sensitive when it comes to performance and thus depend of feedback from real user experiences. One can try to think of hundreds of ways in which performance can be reduced. Only user feedback shows us which ones we need to address.

Kind regards

Jens

I’ll clarify…

I’ve noticed since starting with Open Project, even before the conversion, that there would be stutters of slowness. Things would be moving along fine then all of a sudden a click will take 10 seconds or so to come back. Now after the conversion, those 10 seconds are more like 30 to 50. These do not happen on a regular schedule but when they happen they are very noticeable.

It’s not any particular action, but rather when it happens it happens for all actions and for all users. The system it’s on is responsive though as it’s also running the Postgres server and I can quickly return queries while Open Project is hung up. I don’t know if there is any one page causing the problem but once it starts it affects any page.

We have several projects defined in a tree structure with multiple branches, but most of the tasks are the imported tasks and they are all into one project. In that project they are about 99% of the 10k tasks.

Only about 5% of the packages are related to each other in any way and most of those are related tasks, not parent/child.

As to the environment, I’m on Windows 2008 R2 server, 4GB RAM and Xeon CPU @ 2.4GHz. Drives are just standard SATA 7200 RPM drives. I don’t notice any CPU spikes or memory spikes during the slowdowns.