One of 2016's key buzzwords was serverless - but is anyone sure what that means? Both AWS Lambda and Kubernetes Jobs provide their own solutions in this space, but..
I'm not sure that we have a clearly defined standard on what serverless should look like and how it should behave.
Here are some of my thoughts on serverless based upon what I've seen this year:
- tends involve invoking short-lived functions (Lambda has a default 1-sec timeout)
- does not publish TCP services - often accesses third-party services or resources
- tends to be ad-hoc/event-driven such as responding to webhooks
- should deal gracefully with spikes in traffic