Fork me on GitHub
#onyx
<
2017-05-09
>
jeroenvandijk07:05:05

@camechis last time we used it (less than a year ago) there was a global queue per account. Meaning that if you have one function that has a backlog, it means all functions have a backlog and won’t work. Don’t use it for anything serious I would say

jasonbell07:05:18

@camechis AWS Lambda is nice but I do worry about the long term costs as you’re billed per ms usage against the lambda execution. I’m using it on Kinesis pipelines at the moment (but it’s all in Python).

lmergen15:05:06

also, keep in mind that if your instances fluctuate in load, it needs time to "warm up" to adjust for increased loads

georgek19:05:29

@camechis I’ve used it for event notification with elastic transcoder. The main thing I found frustrating was the size restrictions on the backing jar. You get, I believe, around 10MB per lambda. That forced very careful management of dependent libs, etc, and made some things I wanted to do in a lambda via clj basically impossible. Of course, that was my first lambda and transpiling cljs down to minified js seemed like an alternative that would potentially play much nicer with the size restrictions.

Travis19:05:51

thats good to know