important instance shit

149 readers

1 users here now

founded 1 year ago

MODERATORS

[email protected]

infra: email notifications might be a bit spotty (awful.systems)

submitted 2 months ago by [email protected] to c/[email protected]

14 comments fedilink hide all child comments

we’ve exceeded the usage tier for our email sending API today (and they kindly didn’t email me to tell me that was the case until we were 300% over), so email notifications might be a bit spotty/non-working for a little bit. I’m working on figuring out what we should migrate to — I’m leaning towards AWS SES as by far the cheapest option, though I’m no Amazon fan and I’m open to other options as long as they’ve got an option to send with SMTP

you are viewing a single comment's thread
view the rest of the comments

[–] [email protected] 4 points 2 months ago* (last edited 2 months ago) (1 children)

so that seems to be a thing it doesn't in fact do! I mean, you'd think it would

[–] [email protected] 3 points 2 months ago (1 children)

nah, I wouldn't (think that). I've read some of the lemmy code.

[–] [email protected] 5 points 2 months ago (1 children)

so do you want to hear something terrible? if you give Lemmy SMTP credentials for notifications but it can’t auth against the server for any reason, it’ll just hang the entire lemmy backend forever

which makes switching email providers a bit nerve-wracking, you understand

[–] [email protected] 4 points 2 months ago (1 children)

okay uh don't hate me too much for this but: second service instance postfix, which just acts as a forwarder to whatever you pick? that can have static auth in it for lemmy and thus break as little as possible, and it'll just always queue flush

basically ye olde smarthost, SaaS nightmare edition

[–] [email protected] 3 points 2 months ago (1 children)

heck it can run on the same actual box, postfix is not heavy

[–] [email protected] 3 points 2 months ago (1 children)

yep, literally just a side service in the same flake definition. only downside is you'll need to feed logs to somewhere, and monitor queue depth. but those are both trivial (and you could even just trigger for queue depth over n for $time, because that's enough indication of work/fail)

[–] [email protected] 4 points 2 months ago (1 children)

huh, I kind of like it

we’ll see how the migration goes in staging — if it’s smooth, I’ll add postfix as a todo. if there’s a bunch of problems, postfix might be the most immediate way out

[–] [email protected] 4 points 2 months ago

it's just the right amount of jank to be a perfect solution