AWS Lambda function state (zappa)

Question

I think to optimize flask app lambda server by adding internal cache for relatively slow to change data (e.g. site dropdowns might changes few times per year). I use zappa to deploy to lambda. Does it make any sense? Or does it flash memory each time a request processed. I know that I can not rely on aws preserving state, my goal here optimize performance a bit without spending a fortune on some redis instance not to mention ElastiCache.

UPDATE: Yup, the serverless deployment frameworks like zappa recycle the state, so why I should not. Below a hackernoon blog discuss the state recycling in greater details

https://hackernoon.com/write-recursive-aws-lambda-functions-the-right-way-4a4b5ae633b6

Whilst Lambda functions are ephemeral by design, containers are still reused for optimization which means you can still leverage in-memory states that are persisted through invocations.

Not sure can one invalidate such cache, env variables are likely local to lambda instance, http, sns probably difficult/expensive.

Answer 1

Yeah, that's not gonna work with Lambda.

You have to use some sort of 3rd party cache.

---

If caching only your GET requests is good enough for you, you could use a CDN for that.

I personally use CloudFlare CDN that caches all GET requests for n minutes. And you get a lot of requests for free. You just have to define a custom Page Rule to cache everything for a certain URL pattern.

Of course, same thing can be done with CloudFront (to stay within AWS ecosystem) or probably most other CDNs.