The YAML vampires are eating the Telegram bot and I got a warning that it will be removed. So I tried to use the new one instead. Apparently, my configured notifier was migrated, and I tested using developer tools, this works.
No. You can see in the error message that it tries to call notify.telegram_bot_xxxx_yyyy. If I add another notify., it just gets even worse. Then this is the error message:
Failed to call notify.notify.telegram_bot_xxxx_yyyy, retrying at next notification interval
Now even the direct message sending through an action is broken. When I go to my Telegram integration, it says there is a timeout. Is it possible that the Telegram service is down?
Once I can get back to the action working, I’ll file a bug for the alert notifier.
OK, seriously, Home Assistant does something very wrong. I entered Hass’s container and tried out the instance of apprise installed by Hass, and it works. But for some reason it doesn’t work from Hass itself.
I don’t know what I did, but now it works. After trying Apprise on the container, it started working. Now it works even after I recreate the container.
The initial feedback from the PR indicates that having alerts be broken for all telegram users is the recommended path forward - and then we can each replace alerts with custom automations.
@PeteRage thanks for sharing what is happening. Noticed recently that all my alerts: were able to send email but no longer worked with telegram. Fortunately I found this during the a file alarm test and not in a real emergency.
Have you found a blueprint or script that gives the basic alert: functionality and works with multiple telegram targets? The one here,
apparently only works with the single notification target. If you want to notify your entire family, you are out of luck. Such a pity that the simple but useful alert: integration was left to bit-rot with no replacement.
Thanks @PeteRage, @petersohn for the hints. Both look interesting. For what is worth, yesterday I migrated to the blueprint, for 16 alerts, I went from some 100-150 lines of yaml to 500, not cool, and you loose some flexibility (single retry timer instead of a list of of times), need to use a script hook to notify multiple people etc.
FWIW I ended up following @PeteRage idea and migrated again to Alert2. Smaller userbase but really well done. The yaml syntax ended up being more compact and easier to read than the original alert integration and very flexible. Comes with a nice UI card that can be used to snooze / ack alerts.