Please let me know your best practices on how to accelerate scanning time.
We connected SonarQube (self-hosted) to pull requests
Every time there’s a new pull request it takes 25 minutes - it’s very slow and people complain about it.
I wish we could scan the changes (delta) only - is this possible at all ?
If not how to shorten scanning time? Shouldn’t we run it on ALL pull requests?
We use SonarQube 8 (self-hosted)
We scan Java
Welcome to the community!
First, your version is past EOL. You should upgrade to either the latest version or the current LTS at your earliest convenience. Your upgrade path is:
8 → 8.9.2 → 9.0.1 (last step optional)
Regarding your actual question, I’m pretty sure (without actually going back and checking the release notes ) that there have been some speed improvements across multiple languages between then and now. So just upgrading should help.
You don’t mention what language(s) you’re analyzing, but for C and C++ you have the ability to use caching and turn on multi-threading.
In general, speed of analysis is a topic for us in the next year or so, so even beyond the improvements you’ll see after upgrade this should get better “soon”.
I’m talking about Java
First of all I can only emphasis Ann’s suggestion. Many speed improvements have been done across multiple languages and you would benefit of it while upgrading.
Scanning only the delta changes for Java is not yet available. However additionally to the SonarQube’s upgrade, if you are not currently doing it, you could modify the configuration .yaml file to enable the cache…
Lastly as we are going to improve the speed of analysis, would you please let us know how many modules, .pom hierarchy and files per module your project contains. Thanks