There is a logical issue since the beginning of the cronmanager implementation in #306

It only affects installations which are using docker and keep the base_url at localhost and a different port then 80. It's quite simple (e.g. $config['base_url'] = 'http://localhost:8086/';)

localhost inside the container !== localhost in the host machine.

Reason is that we use port 8086 per default in our docker-compose.yml. This leads into the problem that the mastercron is correctly called inside the container with http://localhost[:80] (see Dockerfile) but the cron/run method gets http://localhost:8086 from the base_url config (via local_url() method). This obvisouly fails inside the container as in this little cute blue whale Wavelog is running on port 80, and not 8086.

There was a good solution around a year ago by @winnieXY in #795 but the people (and myself) forget about the config option. Therefore we use now http://localhost[:80] by default in docker environments which always works (because inside the container Wavelog always runs on localhost port 80). Thanks to @blhoward2 for poking me here. Even this is not a bug, it leads into a bad user experience when people running Wavelog in docker for the first time on their local machine and the cronmanager fails...

Solves

• All cron jobs are failing #2208
• LOTW cron job always fails, but manual sync works #2180
• .. and probably some more