Update: I learned about `aws cloudformation package`. You don't even need to know how to write bash to upload to s3 :P http://docs.aws.amazon.com/cli/latest/reference/cloudformation/package.html …https://twitter.com/southpolesteve/status/943935246009020419 …
-
-
very tempting! would love a tool that is a watchified version of aws-cli deploy that lets you opt out of CF’s slowness through API calls (for lambda UpdateFunction, for example), but then again there’s this sage advice:https://twitter.com/southpolesteve/status/943943227438137345 …
-
Yes to both these things; haven’t got the .arc to sam baked yet but I think it’ll be more of an eject option rather primary workflow. CF has a bunch of tradeoffs.
-
what tradeoffs remain if you use CF’s schema but not CF’s implementation? might help tool interop to have the same target output and separate the actual deploy side effects.
-
That seems pretty doable using cf changesets. If the changset is code only do it via apis
-
i’ve found changesets can be slow. the idea with a watchify approach is that the diff can happen locally.
-
some work-in-progress numbers; I can generate a hello world api gateway/lambda combo per an .arc file with just one route `get /` which desugars to 600+ lines of yaml in 64ms. a CF changeset deploy takes a minute. scripted arc `npm run deploy` takes 3s max. /shrugpic.twitter.com/kiqRLv6cza
-
(aside: its still so fkn cool to me that we are measuring zero interruption/downtime deployment times in seconds and minutes tho)
- 4 more replies
New conversation -
-
-
cloud formation is great but ya can be slow sometimes. Custom resources in cloud formation are crazy powerful
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
API calls are way faster but the amount of maintenance on keeping that up to date is… Non trivial

- 1 more reply
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.