post-fail hook

[greenship24]

Looking at the new hooks (and v5 is great, thank you guys), why is there no post-fail hook?

Seems:

• pre is for preparation (in case something might change)
• deploy (success in changing something)
• post (regardless of outcome)

To me it makes more sense to split post into post-success and post-fail. Post-success triggers on exit 0 and post-fail triggers on anything other than exit 0 with some preconfigured (and passed) env var to the post-fail hook.

I understand you could do something akin to using pre and post together (a pre without a post notifies you something is wrong) but still seems a bit hacky instead of two clearly defined post hooks. You could argue post-success might be unnecessary with the deploy hook but I guess it would notify in case of successful run even if a cert is not renewed.

[ldez]

Hello,

v5 is great, thank you guys

Thank you ❤️

---

could you explain what is the goal of a post-fail hook? I mean what to do want to automate in case of error only?

From me, the main usage of hooks is something like that:

• pre: start or stop a process
• deploy: copy the certificate somewhere else
• post: stop or restart the process related to pre

Note that lego is not reacting to exit codes (this is not a shell script).

[greenship24]

I guess I was looking to be notified if something had gone wrong in the program itself (and sending the exit code to me automatically.) I know this can be accomplished using cron anyhow and simple bash. That and monitoring the cert expiration should notify me.

I think my mind immediately went to something like restic where I can get a notification on fail with an included exit code so I can go fix it up before it becomes a problem. Perhaps I thought too much (or not enough) and didn't think to just use the simpler tools I've already got.