Turning on debug mode will make the exceptions show in the browser. This is useful to debug issues in the web interface, but not suitable for production environment as it has performance consequences and might leak private data.
Weblate can produce detailed logs of what is going in the background. In
the default configuration it uses syslog and that makes the log appear either in
/var/log/syslog (depending on your syslog
The Celery process (see Celery を使用するバックグラウンド タスク) usually produces own logs as well. The
example system-wide setups log to several files under
Docker containers log to their output (as usual in the Docker world), so
you can look at the logs using
Not processing background tasks¶
Lot of things happen in background Celery workers. In case things like sending out e-mails or component removal does not work, there might be some issue with it.
Not receiving e-mails from Weblate¶
You can verify whether outgoing e-mail is working correctly by using the
sendtestemail management command (see Invoking management commands
for instructions on how to invoke it in different environments) or using
管理画面 under the Tools tab.
These send e-mail directly, so this verifies that your SMTP configuration is correct (see メール送信の設定). Most of the e-mails from Weblate are however sent in the background and there might be some issues with Celery involved as well, please see Not processing background tasks for debugging that.
Analyzing application crashes¶
In case the application crashes, it is useful to collect as much info about the crash as possible. The easiest way to achieve this is by using third-party services which can collect such info automatically. You can find info on how to set this up in エラー レポートの収集.
Lots of tasks are offloaded to Celery for background processing. Failures are not shown in the user interface, but appear in the Celery logs. Configuring エラー レポートの収集 helps you to notice such failures easier.
In case Weblate performs badly in some situation, please collect the relevant logs showing the issue, and anything that might help figuring out where the code might be improved.