Obvious choice is python as there are already mature web libraries like flask, but my new favorite stack is using lambdas like a normal server. If you go that route you can write them in Java, Go, PowerShell, Node. js, C#, Python, and Ruby. I'm working on a project right now, I wrote the server in node as an (aws) api gateway lambda proxy, without using any libraries. I eventually decided to rewrite it as a node/express server, and it's kind-of badass. The most cost efficient option I can imagine while also effortlessly scaling to meet extremely high demand. Lambda can scale up to 1000 concurrent connections, in theory you would need 125 8 core ec2 boxes to match the "multi-threaded" processing power. That's a pretty bloated estimate as there are other variables that will slow down lambdas, but you get the idea. The second reason I love this system is deployment bliss. It's fast to launch stacks with cloudformation when the heaviest thing it has to do is update a single lambda. This means you can build continuous delivery/integration automation into your dev tooling (github actions), you can quickly launch "staging" stacks, removing the need for a blue/green type system. Updates are immediate and rolling, meaning you don't have to launch the new task before stopping the old task to achieve zero downtime. Ok I'm done geeking out on my new serverless server design.
I'm not getting in the middle of a python vs ruby debate, mostly because I don't care lol. There is nothing wrong with either language but devs will always swear their language is best because it worked well for their specific need.
Wow, we got a Lambda lover here! 😀 Happy to hear that you found a stack that feels right to you! I've never quite looked into AWS before, but I probably should.
Ha. I've deployed aws ecs in production for years, it really depends on your apps needs. It's been my experience to make a good serverless app you have to write a lot of your own devops tooling. But once that's in place, the serverless nature alleviates all the hardest pain points surrounding scaling, versioning, and zero downtime deployment infrastructure.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
Obvious choice is python as there are already mature web libraries like flask, but my new favorite stack is using lambdas like a normal server. If you go that route you can write them in Java, Go, PowerShell, Node. js, C#, Python, and Ruby. I'm working on a project right now, I wrote the server in node as an (aws) api gateway lambda proxy, without using any libraries. I eventually decided to rewrite it as a node/express server, and it's kind-of badass. The most cost efficient option I can imagine while also effortlessly scaling to meet extremely high demand. Lambda can scale up to 1000 concurrent connections, in theory you would need 125 8 core ec2 boxes to match the "multi-threaded" processing power. That's a pretty bloated estimate as there are other variables that will slow down lambdas, but you get the idea. The second reason I love this system is deployment bliss. It's fast to launch stacks with cloudformation when the heaviest thing it has to do is update a single lambda. This means you can build continuous delivery/integration automation into your dev tooling (github actions), you can quickly launch "staging" stacks, removing the need for a blue/green type system. Updates are immediate and rolling, meaning you don't have to launch the new task before stopping the old task to achieve zero downtime. Ok I'm done geeking out on my new serverless server design.
Python is half as elegant as ruby with twice the overhead, and less shiny
This is wrong on all 3 points lol
I'm not getting in the middle of a python vs ruby debate, mostly because I don't care lol. There is nothing wrong with either language but devs will always swear their language is best because it worked well for their specific need.
Wow, we got a Lambda lover here! 😀 Happy to hear that you found a stack that feels right to you! I've never quite looked into AWS before, but I probably should.
Ha. I've deployed aws ecs in production for years, it really depends on your apps needs. It's been my experience to make a good serverless app you have to write a lot of your own devops tooling. But once that's in place, the serverless nature alleviates all the hardest pain points surrounding scaling, versioning, and zero downtime deployment infrastructure.