In this page

Configuring app logging

When investigating a run-time problem, you may need to increase the logging level for the app. Doing so, the app writes more run-time details to the Jira log which helps us understand what's going on.

To increase the logging level via the Bitbucket REST API without stopping Bitbucket:

  1. Open an operating system console (command-line).
  2. Execute this command (using the proper parameters):
    curl -u <admin_username>:<admin_password> -v -X PUT -d "" -H "Content-Type: application/json" <bitbucket_base_url>/rest/api/latest/logs/logger/com.midori/debug
  3. Check if a line like this appears in the Bitbucket log:
    2019-09-10 16:15:31,768 WARN  [hz.hazelcast.event-3]  com.midori Switching to log level [debug]

Alternatively, to increase the logging level permanently (i.e. even if Bitbucket gets restarted later):

  1. Stop Bitbucket.
  2. Edit the $BITBUCKET_HOME_DIRECTORY/shared/ file (create the file if it doesn't exist yet).
  3. Add the following line and save:
  4. Start Bitbucket.

Now execute the app function that you think fails, and check your Bitbucket log to see the details.

If you can't understand the log lines, report what you have found to our support. We are there to help you.

Typical problems and their solutions

My push is rejected with the message "Too many commits found, retry with smaller data! (max. 1000 accepted in one batch)"

You're trying to push (or merge) an unexpectedly large number of commits in one go. The app is designed to strictly verify every single commit before accepting it. In order to keep the server stable, it imposes a limit of 1000 for the number of commits that are processed in one push.

The solution is simply splitting the "mega-push" into multiple smaller pushes, each containing max. 1000 commits. As explained in this article, Git allows pushing commits up to a certain commit (defined by its hash). If you have e.g. 2500 commits to push, you can use this method to split those to 3 separate batches.

In the following section, we give a step-by-step guide for this. Please note although that the steps are explained with the Git commandline client, it is possible to adapt these to your favorite Git client, too. It is just super-easy with the commandline client.


  1. Determine the hash of the 1000th commit to push by using this command (Linux/bash syntax):
    git --no-pager log <REMOTE>/<REMOTE_BRANCH>..<LOCAL_BRANCH> --oneline --reverse | head -n 1000 | tail -n 1
    For example, if you try to push from your local master branch to your origin's master branch:
    git --no-pager log origin/master..master --oneline --reverse | head -n 1000 | tail -n 1
    It will show the hash of the 1000th commit, for example:
    108a04ca Fix for the killer bug FOO-123
  2. Use this commit hash to tell Git up to which commits to push:
    In our case:
    git push origin 108a04ca:master
  3. Push the "remaining" commits simply by:
    git push
    If it is rejected because you still have more than 1000 commits, then go to step 1 to push the next batch.
    If it is accepted, then you are done. Congratulations!

My push is rejected with the message "Access key "Access Key User -" is not allowed to push commits."

You're trying to push commits using an SSH access key set up for system use, while Better Commit Policy for Bitbucket is configured to reject all commits pushed by external systems.

See the corresponding configuration guide.

I see an error message "An error occurred while processing the request" at pull requests.

Secondary symptoms: the spinner of the "Merge" button does not stop and an exception was thrown in PullRequestService.canMerge().

This is caused by a minor Bitbucket bug (BSERV-10045) which was fixed in Bitbucket 5.4.0.

You can safely ignore the red error message: it is harmless and it is displayed only once. Just reload the page in your browser to make it disappear. And then as an actual solution, upgrade to Bitbucket 5.4.0+.


Ask us any time.