Number of issues in a project before performance degredation

[Emre Toptancı _OBSS_]

I know there are several questions and discussions on this community about max number of issues allowed in a single JIRA project and the simple answer is "there is no practical limit"

My question is somehow different. ACP-100 JIRA Administration Certification Exam Prep Course mentions online and offline archiving. It says Online Archiving is the practice of moving JIRA issues to another JIRA project on the same system, with the aim of improving search performance. The course material says Online Archiving is recommended because having too many issues in a single project might decrease search performance. It makes some sense but does not say what "too many" is, neither does JIRA documentation. 

So my question is: Based on your experience what is the number of issues in a project before search performance starts to drop noticably? 100K, 1M, 5M ?

[John Gill]

I wish we had some more clarity on this question, specifically in a cloud environment. 

[somethingblue]

Hi Emre,

There may be more specific information out there, however, from a support perspective I have seen this is dependent on the resources your system has.  I have seen instances with a million+ issues running on servers that were optimized by one of the partners that run without issue. 

That being said, JIRA's performance can usually be tied back to resources on the machine running JIRA.  The JIRA Sizing Guide is a good place to start in regards to the resources JIRA needs up to the point it's recommended to have someone come onsite to take a look at your deplyment and your needs!

Cheers,

Branden

[Emre Toptancı _OBSS_]

Hi Branden, 

We also have a few customers with close to a million issues and with careful configuration they are working relatively fine.

The sizing guide you mentioned (or any other JIRA performance optimization resource) only focuses on number of projects, custom fields, issues, schemes, etc. Since exam prep material explicitly points out the performance cost of having "too many" issues in a single project, I am trying to get some insight into this specific topic.

I am hoping someone that has faced a significant performance problem and mitigated the problem with dividing projects will contribute to this question and give us numbers.