As time passes, repositories tend to have larger amounts of data - such as pull requests and commits. This slows down the DVCS sync and makes Jira poll for more data from the repositories.

The suggestion is to limit the data to some time back (like 1 year) and create a flag to tune this, just like Safeguards for Jira's index.

Older data could only pulled on initial/full syncs - maybe it could have a different flag that limits the full sync too when set.