News for hackers. Stuff that matters.

A comment on How we spent $30k in Firebase in less than 72 hours. Parent comment

Re:How we spent $30k in Firebase in less than 72 hours by onion2k on Thursday August 02, 2018 @01:52AM (#17669475)
Your app shouldn't just fail if an API is unavailable. You should code things to fail gracefully. In the case of a traffic spike replacing the front end with an email capture form and a message saying "things are really popular right now, how about we remind you about this tomorrow (with a discount code!)" works well.

This is hard for some things, but your startup failing because you didn't want to do it is much harder in the end.

[ Parent | on HN ]
Re:How we spent $30k in Firebase in less than 72 hours by asadjb on Thursday August 02, 2018 @03:03AM (#17669698)
Most applications have atleast 1 SPOF, the DB. Just because the DB is now the Firebase API doesn't change anything.

It would be very difficult to build products in a reasonable amount of time if everything has to be coded defensively. I can build my app quicker, and remain sane, if I assume that the DB will always be available, and just fail if the DB isn't there. Same for things like S3 (which I think had only 1 large scale failure in recent history), Redis, etc...

There are APIs which can be unavailable and you need to work around those. For me, these are mostly third party services that I don't control. But then again, I'm not building the next Netflix. I don't have enough engineers to build an app that works with a chaos monkey!

Not the best approach for all applications, but has been good enough for most projects I've worked on. Just my 2 cents.

[ Parent | on HN ]