Summary
We have one or two time a weeks a hole in our visit log (table matomo_log_visit). Every time it start around midnight and ends one or two day later around midnight or 1am. So the hole is almost 24 or 48 hours.
Steps to reproduce
We are using Matomo 4.13.2 in docker. Database 10.9.4-MariaDB. PHP 8.1.14
Expected behavior
We should not have any hole in our visit logs.
Observed behavior
The queue tracking plugin is on and receive events in the redis. Their is no database backup during this time or any interruption with the database or the queue system.
Every times it happens we need to replay logs from our haproxy access logs and it’s very time consuming.
I don’t see from where it can come from, the system works well for some day and stop working for 24-48 hours straight. It’s very weird…
1 post - 1 participant