Loading project page takes a long time

[Iona Augustine]

Hi,

We have issue of taking a long time when loading project pages and just project pages. It takes 10-18 minutes, depending on how big the project is.

The server and database server performance are decent, if not, everything should be slow as well.

Has anyone here ever experience this as well?

Thanks!

[Jason U]

Hi @Iona Augustine,

Long load times on Fisheye/Crucible project pages are usually related to indexing, database performance, or JVM/OS limits. Based on Atlassian’s documentation, the main areas to check are:

1) Repository size and indexing load
Large repositories or many branches/tags can slow project pages significantly.
Ref:

https://confluence.atlassian.com/fisheye/improve-fisheye-scan-performance-960155798.html

2) Database indexing
Missing/insufficient DB indexes can cause long query times when loading project metadata.
Ref: https://confluence.atlassian.com/fisheye/tuning-fisheye-performance-960155706.html

3) JVM heap and file descriptor limits
Low heap or OS file limits often lead to delays during page rendering.
Ref: https://confluence.atlassian.com/fisheye/tuning-fisheye-performance-960155706.html

4) Storage location
If repositories are stored on remote/NFS locations, Fisheye response time slows down.
Ref: same as above.

Quick checks you can run now:
• Review server CPU/disk activity during the slow load
• Check repository sizes and indexing logs
• Verify JVM memory settings
• Ensure DB indexes match the recommended configuration

[Mia Tamm _Simpleasyty_]

Hi @Iona Augustine

Why it’s slow

1. Unindexed database queries – Looking through millions of rows without an index can take minutes.
   
   
2. No caching – Every page load rebuilds everything from scratch.
   
   
3. Too much data at once – Returning thousands of items in one go is slow to serialize and send.

How to fix it

1. Add appropriate indexes on the columns your queries filter or sort by.
   
   
2. Introduce a cache (Redis, Memcached or your framework’s cache) so repeat views are fast.
   
   
3. Paginate or lazy‑load large lists so each page only handles a small chunk.

First step: Turn on simple timing (or a profiler), see whether the database, the cache layer, or the network is the bottleneck, then apply one of the above fixes.

— Mia Tamm from Simpleasyty