Feeling happy to share one of my projects where we reduced AWS bills by over 45% since Feb, 2020.
Haven't even started with optimizing reservations yet. See d reservation bar falling every month(first day of each month).
💰🐌🪲🪱🐍💰
#CloudComputing#100DaysofCode#cloudcosts
A random thought -
Most of AWS services are priced on Compute, RAM and Disk.
DynamoDB is different and is priced on number of requests it actually serves (and disk).
What this means is that if you write very sluggish queries that take a lot of time to execute, AWS takes the hit on extra compute.
Let's say, there are two users making 100k requests a min. First one writes good queries which run at 5ms average latency while the second one takes 10ms on an average to run a query. Both pay same to AWS. Kind of free compute to the second guy. 🤯
First Valkey (30% cheaper than redis), and now plain price reductions in DynamoDB and KeySpaces.
Very unusual of AWS to reduce prices on their core services like this.
I hope next in line will be the archaic data transfer charges. 😆
@awscloud
Last week was action packed with AWS rolling out several big changes to its pricing. The limelight was however stolen by DynamoDB and KeySpaces. The prices for DynamoDB on-demand throughput were reduced by 50% and global tables by up to 67%. Along with this, Amazon Keyspaces also announced price reduction by up to 75%.
There have been several similar announcements made by AWS in last few weeks and there is a clear pattern emerging from all of these. A few weeks ago, AWS Elasticache launched Valkey and it was priced almost 30% cheaper than the Redis counterpart. A couple of weeks ago, AWS made all CloudFront requests blocked by AWS WAF free of charge. Then came the news where AWS waived off subscription fee per configured user and couple of other costs in AWS DataZone service.
All of these announcements are unlike the usual cost optimisation avenues that AWS usually opens up by technologically advancing its products and passing on some discounts to the customers. These are plain price reductions and probably hints at a bigger price war brewing between the titans of cloud computing. We should be ready for more such reductions in near future.
Regardless of how things pan out for AWS and its impact on competitors, it’s a great news for the customers.
I send out a weekly newsletter on latest news around AWS cost optimisation and random banter.
https://t.co/E4HMLnASEG
Top 5 ways to reduce your AWS costs organically.
🗑️ Delete unused resources: Let go of what isn't needed.
↔️ Right Sizing: Get back in shape
❄️ Spot Instances: Leverage the scale of cloud
✤ Auto Scaling: Leverage the elasticity of cloud
🐰 Application Performance: Improve your code
Full post here --> https://t.co/QTwA7kSGTx
Its wild how much you can cut your AWS spends by designing stateless applications. That too without buying any savings plan or anything.
💰💰💰💰
Two ways where going stateless can save you literally tons of money.
1. Easier to use spot instances. Spot instances can be upto 90% cheaper than regular on demand ones.
2. Easier to autoscale dynamically. Since there is no state on the VM, you can terminate and spin up instances whenever you want to without any side effects.
AWS' data transfer costs are a relic of the past.
Bandwidth costs have plummeted by over 90% in last 15 years but it has never occurred to @awscloud to rethink their exorbitant data transfer costs.
Linode is literally 10x cheaper than AWS on this.
C series instances are cheaper than M series instances, which in turn are cheaper than R series instances. Graviton instances are cheaper than AMD which in turn are cheaper than the Intel ones.
You need to select the instance family that can run your workload with minimal cost.
AWS Cost Optimisation Tip: Migrate your ElastiCache workload to ValKey engine with Serverless priced 33% lower and node-based priced 20% lower than the Redis engine.
Go to https://t.co/5BkaN6LTso for more such tips.
We're hiring a backend software engineer to help build our cost management product. Very small team, so lots of ownership over how to build things. Hybrid in-office (3x/week in-office) in San Francisco. We'll help with relo.
Buying AWS savings plan with one year commitment actually means you only need to commit for 9 months. You'll hit break even at 9 months.
Go ahead and buy some of it today!!