This page is not created by, affiliated with, or supported by Slack Technologies, Inc.
2021-01-31
Channels
- # babashka (23)
- # beginners (65)
- # calva (126)
- # cider (23)
- # clj-kondo (10)
- # clojure (41)
- # clojure-europe (5)
- # clojure-nl (2)
- # clojure-uk (6)
- # clojurescript (78)
- # contributions-welcome (1)
- # datomic (9)
- # events (1)
- # graalvm (2)
- # meander (10)
- # off-topic (81)
- # pathom (5)
- # shadow-cljs (3)
- # spacemacs (6)
- # vim (3)
- # xtdb (4)
Hey Joe, what is your specific use-case for canary deployment? Also, could you elaborate on what you mean by “plans for serverless “?
Basically by canary deployment, is it possible to do a rolling update in production? if something goes wrong then immediately revert to previously working production.
by Serverless I believe thats what the Lambda proxies solve. All in all Datomic Ions seems to answer my needs.
RE: canary deploy, It depends again on what you mean by “something goes wrong”. Right now, if you initiate a deployment and the ec2 instance fails to complete its code deploy steps, then the deployment is marked as failed and (in a production topology) there is zero downtime. Now, if the code deploy succeeds on a given instance and your code starts up but the latest release of your code is “too slow” as measured by a cloudwatch metric you have set up, datomic ions can’t fail the build on your behalf because it appears to have succeeded as far as Code deploy is concerned. Hope that clarifies things and I’m glad I could help.