TLDR: My dog ran away because I didn’t know about continue_on_error 
Kind of 
Adding my vote for having continue_on_error in the UI, and a background story:
I have an automation that notifies 3 devices (TV, speaker, computer) when my automatic front gate is opened. It notifies my phone instead if I’m not home. The same automation then continues into a while loop that notifies me every minute while the gate is still open.
Today I left on my motorcycle and closed the gate behind me, but because of a Zigbee fluke, the gate opened again. Well, I had an automation just for that—or so I thought.
In the first notification, the automation stopped because the TV was off 
You see, I had already run into this in another automation, which I “solved” (now I know, not really) by running the notifications in parallel. An error in sequential execution can stop the automation, okay. But an error in an asynchronous action shouldn’t, I thought.
Well, in this automation, the reminder loop comes after the first notification, so it was never reached because the first notification couldn’t be delivered to the TV. I wasn’t reminded on my phone about the open gate, and my dog of course didn’t miss the opportunity to go see sniff the world.
Home Assistant has a learning curve, of course, and today I learned about continue_on_error and the Alert integration, which I will try next. Having options like this available in the UI helps make the behavior clearer, easier to predict, and the learning more intuitive—IMO. Finding that I could change the automation modes (single, restart, queued, parallel), for example, was a game changer in my automations.
Of course, I could’ve tested the automation better, and I could have RTFM, but oh well. I’m off to reviewing all my automations and reading about the Alert integration now.
P.S. The dog was found and had some crazy stories to tell.
P.S. 2 Attached image of pain and suffering.
