AutoSpotting VS aws-lambda-extensions

There's a service that handles obtaining the cheapest spot instances to your specs by a former aws engineer https://github.com/LeanerCloud/AutoSpotting

Effectively, it starts up cheap spot instances (based on specified criteria) across a variety of instance types to replace whatever regular instance in an autoscaling group comes online and then spins down the regular instance.

EG: That m4a you wanted may be expensive... but nobody is using m4ad so it's 85% off and it meets the specified CPU/RAM requirements... auto spotting will spin it up instead.

Having used it on and off over the years it is sometimes eyebrow raising to see 4xl boxes running cheaper than the xl box they replaced :)

I've been heavily using ChatGPT on my autospotting.io software, where it 3-5x my productivity, releasing in 6 weeks more improvements than in the previous 6 months combined: https://github.com/LeanerCloud/AutoSpotting/discussions/489

Relevant: https://github.com/cloudutil/AutoSpotting

I've seen some third party services that automate migration to / replacement with spot instances, but haven't used them yet personally.

Going serverless, in many places, has been the most effective cost optimization for me.

I've also been maintaining a Golang library that essentially exposes the same data structure used under the hood by ec2instances.info as a Golang data structure, which is very convenient for embedding into other tools that can do lots of things with this raw data, like I do it in my AutoSpotting project.

I think autospotting can do this.

https://github.com/AutoSpotting/AutoSpotting is an interesting project that might help you with implementing this. It will automatically swap-out on-demand instances for spot and auto-restart as on-demand if spot capacity disappears.

we scale up to about 100 machines. We use spot instances EXTENSIVELY. And that configuration was tricky actually. Its been a couple of months now. Works pretty ok.

k3s is actually pretty simple to use now. the tricky part was to integrate with https://github.com/kubernetes/cloud-provider-aws and https://github.com/DirectXMan12/k8s-prometheus-adapter

The hardest part is to get it to work with spot instances. we use https://github.com/AutoSpotting/AutoSpotting to integrate with it.

This should look familiar to this AWS Repository. I'm still personally exploring more of how I can customize and instrument some of this code, but the extension is registered and ready for use by this function.

There are some extension examples github. I am using go-example-telemetry-api-extension as a starting point. You can see the code in my Lambda-Telemetry-API-Loki repository.

You can't send some of your lambda logs to one log group and some to another without directly invoking the sdk, which you probably don't want to do. Maybe someone has created some lambda layer to handle this, but not one that I know of. I see people have tried to get fluent bit to work in lambda with mixed results https://github.com/aws-samples/aws-lambda-extensions/pull/42.

For even more inspiration related to the Logs API, checkout the AWS-provided samples. Some examples include batching the logs and delivering them to S3 or a template to get you started.