We have built a Telegram workflow, which starts with a Telegram trigger. After some time (sometimes a day, sometimes 3 days, etc.), the Telegram trigger doesn’t respond. Also, the workflow, which is using the current bot credentials, saves and activates/deactivates the workflow, just loading and does not save.
When we refresh the bot token, everything starts working (the trigger is responding, and we can make edits in the workflow).
We know that this is the only workflow that is using the Telegram trigger WITH those credentials.
Also, sometimes, the trigger is not responding, AND after an hour, it runs with all the messages in a row with like 5-20 seconds between them. But this may be a Telegram problem holding our messages for some reason.
As I mentioned above, refreshing the token works, but we need to do this regularly, which is not good.
Also, we are using our own custom base URL in credentials, which is hosting the Telegram API via Google Cloud Run (so we can work with bigger files than the 20MB Telegram default)., Just now, I changed ONLY the Base URL to the default, and the bot started to work
Maybe someone has some experience with problems like these and could point us to what else we could check
Self-hosted n8n 1.120.4, Postgresql DB