This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2016-04-29
Channels
- # admin-announcements (1)
- # aws (10)
- # beginners (76)
- # boot (53)
- # braid-chat (1)
- # cider (80)
- # cljs-edn (3)
- # clojure (65)
- # clojure-belgium (2)
- # clojure-gamedev (2)
- # clojure-nl (3)
- # clojure-poland (1)
- # clojure-russia (39)
- # clojure-uk (14)
- # clojurescript (91)
- # cursive (62)
- # datascript (1)
- # datomic (9)
- # dirac (34)
- # emacs (25)
- # error-message-catalog (8)
- # events (1)
- # hoplon (88)
- # instaparse (1)
- # jobs (2)
- # jobs-discuss (6)
- # lein-figwheel (7)
- # luminus (43)
- # mount (5)
- # off-topic (7)
- # om (28)
- # onyx (61)
- # planck (4)
- # re-frame (27)
- # reagent (3)
- # remote-jobs (2)
- # spacemacs (3)
- # untangled (136)
it depends .. if you need quick startup and short lived apps, then clojurescript lambdas is fine ..... if a little hit on startup and the lambda lives for awhile, and clojure is good
also depends if the libraries you need to use work with clojurescript or if they need java
jvm ... its not surprising
we have a lambda that's called about 1M times/day with an s3 event, all it does is put the event into dynamo
the clj 95th run time was 10ms, the node version is about 20ms
we were being throttled and i thought it was because of clj startup, but really it was dynamo errors and retries causing the lambda to be alive too long
ended up just using the node version because amazon support understood it better
the 10ms of addl overhead amounts to 3$/month in extra costs, which i am OK with