Closed ghost closed 4 years ago
Thanks for submitting this, @k1-hedayati! I think it's a clever solution, but you're right. This is not very production ready. The reason why is that we get into some tricky cases. For example, if one URL is much slower than other ones, it could cause other checks to miss the beat. The ideal way is having each check run in a separate thread. So, one url don't affect other ones.
Yes making checks async is a different beast. In case you have 100 endpoints and each taking 1 second you won't have a good time with this monitoring but I believe this isn't the case for 90% of use cases so it would be safe to use this approach until there is enough budget to make monitoring asynchronous.
Very true! I do agree it's a pretty good intermittent solution and I do appreciate it :)
Hi there, as I would like to get this running at some point: are there some (positive) news? :)
@ogghi sorry, I was on vacation without access to a computer. Also, I've been busy with other projects/work. I'll try at least solve the conflicts and merge this as soon as I can.
@mtakaki thanks a lot! :slightly_smiling_face:
I don't think I can fix this broken tests, because the account is deleted... I'll favor merging #71 instead. Unfortunately, I'll have to close this PR.
It's a quick a dirty solution that just works, so merging this PR isn't expected. But it will satisfy the needs in #9 to some degree and they can merge this PR in their private repositories.