#aws #lambda #ecs #rds
Welcome to day 21 of the unofficial AWS Cost Optimisation Advent Calendar 2024, where every day we will be sharing new tips or tricks to help you optimise your cloud costs before Christmas 2024.
Today we are going to talk about Serverless and the cost benefits.
Serverless, as the name suggests, is a way to run your applications without the need for servers or containers constantly running. Instead, you only pay for the time your code is running.
So the question for today is "Do you even need a server?".
Some of the most common are:
When people talk about Serverless in AWS, they are usually referring to Lambda. Lambda allows you to run code without provisioning or managing servers where it is essentially functions that have been written waiting to be called.
You only pay when they are actually called.
Some companies write their entire application using only Lambda. These could be entire web apps or APIs.
In this way, if no one is using the application or website, you are not paying for it to be up and running.
It is possible to "lift and shift" existing applications to Lambda but this takes quite a bit of consideration as to the trade offs and complexity. Cost in these cases is just one factor of many before making a decision to move to Serverless.
There are entire papers written on the economics of serverless and whether it is actually cheaper than running servers.
Today we want to focus more on the simpler use cases where you can start to benefit from Serverless with a clear saving.
One of the most common use cases for Lambda is to create microservices. These can be small pieces of code that do one thing and do it well.
Some good use cases of when these might make sense to run in Lambda are if they will only be called infrequently or in sporadic bursts.
Instead of having these as containers or ec2 instances running all the time, you could build them in Lambda using something like the Serverless Framework.
In this way, you are never paying for the server or container to be running for this service when it is not being used.
In addition if you have sporadic workloads that are resulting in having to increase server capacity in your existing infrastructure, splitting these out into a microservice could allow you to scale these independently and downsize your main server capacity.
Aurora Serverless is a way to run your database without having to provision a database server. Instead, you only pay for the time your database is actually being used.
It will grow and shrink based on the demand.
Previously, there was no ability to scale down to zero, but this has been available since November 2024.
This means if you have an application that is used infrequently, you can be paying literally nothing to have it readily available when someone does use it.
Aurora Serverless is compatible with MySQL and PostgreSQL.
If you have a static website, you can host it in S3 and serve it using CloudFront for a fraction of the cost of running servers.
Many CMS systems can be configured to output static websites which can then be uploaded to S3 and avoid the need for server hosting.
One of the considerations when using Lambda is the time it takes to run. Lambda's pricing model is based on memory per second used which means any seconds you can shave off or any memory you can reduce will save you money.
This means it is worth investigating if you can reduce the time or memory taken to run your Lambda functions.
Using the Cost Explorer you will be able to drill down into the cost per individual functions to see which ones could use some work.
Serverless can be a great way to save money on your cloud costs. It is not a sure bet as the economics of Serverless can be complex, but there can be a number of clear wins when looking at the right use cases.
When looking at your workloads today, can you spot any clear cut candidates for Serverless?
To be the first to know when the next edition is published, follow us on LinkedIn, X, BlueSky or subscribe to the RSS feed.
Join our newsletter for Cost Optimization tips and tricks
By subscribing you agree to our Privacy Policy and provide consent to receive updates from our company.