The serverless approach to testing is different and may actually be easier
Paul Johnston

Reading through this two thoughts came to mind:

  1. The old chestnut, ‘everybody has a test environment, some are lucky enough to have a separate production environment’, and you make the point that it’s relatively simple to clone an environment, but not so easy to give it a good work out, especially if you have event driven mousetraps that might have race conditions or similar in them under load.
  2. How do you do canarying in Serverless? This was (reasonably) well understood in the world of EC2 and ELB, but I’m not sure how to gain fine grained control over load balancing within a more service provider managed environment (unless you write your own proxy to front different versions of functions, which opens a massive can or worms, and likely dumps all over any existing CI/CD).