Retrying notify handlers on subsequent runs?

Yup, it's been brought up before -- a long time ago, actually, and
then it wasn't talked about for a while.

I don't think ansible runs should keep state, but there probably needs
to be a way such that if a handler is tagged and you run the tag, the
handler runs.

Are there other ways that people can think of, in terms of
playbook/CLI interface (not implementation so much) that seem
intuitive?

(What Would Other Config Management Tool Do?)

If I’m understanding correctly… I wanted something similar to this the other day. Automatically noticing handler failed and run it “next time” is not good, state and other reasons. But, once seeing handler failure I would like ability to simply rerun handler. It should be very explicit. An option to ansible-playbook similar to tags.

–trigger [,<handler name, …]

It triggers the named handler from the playbook and no other tasks are run. For simplicity, --trigger and --tags are mutually exclusive.

I'd probably say it should be --handler, assuming it didn't run any
non-handler tasks and only ran handlers.

Reasonable?