I’m seeing a weird issue when some of the cron jobs simply don’t execute and get stuck, hanging until the timeout feature kicks in in 12 hours.
Just to be safe, in my code I did add functionality to kill the process after 5 minutes of inactivity, but that code doesn’t seem to be executed either, which makes it seem that there’s something wrong with some of the machines.
In fact, I’m pretty sure it’s some of the hosts, because when the timeouts happen I’m not even seeing the “Detected Node version” and “Starting service” messages in the logs. Logs are showing nothing.
Wondering if anyone else saw this and how they went about fixing it.
This happens at least once a day, on multiple cron jobs, and they’re all hosted in Frankfurt.
One clarification, when you mention that this happens
on multiple cron jobs
are you talking about multiple cron services or multiple runs of one service?
I did add functionality to kill the process after 5 minutes of inactivity, but that code doesn’t seem to be executed either, which makes it seem that there’s something wrong with some of the machines.
I do see some TimeoutError: Navigation timeout of 30000 ms exceeded messages in your logs. I’m not aware of a Render error like that, could you confirm that’s the 5-minute timeout you implemented?
If you’re still seeing regular canceled jobs could you send me the impacted service ids (i.e. crn-...) and approximate times in UTC of the failures (note the times in the Dashboard are local)?
Thanks for replying! I dag a bit deeper and it looks like there is an incompatibility between the “discord.js” package and the servers in Frankfurt. More specifically, they’re hanging at the line that logs in the discord client (which is why I was seeing no output for my cron jobs). This works sporadically, but the timeout occurs 95%+ of the times.
Other areas seem to be fine (I’ve tried Singapore and Ohio).
