The last organization I worked for we were actively trying to reduce the company's AWS costs over a one year period. A few things I've learned from that experience was that VPC endpoints do help to reduce costs. Although you have to pay for them but by utilizing VPC endpoints, we can reduce egress traffic costs to S3 and improve security by keeping your data transfers within the AWS network, while also benefiting from reduced latency since the traffic stays within the AWS infrastructure. We recommended that the company use reserved EC2 instances, got rid of EFS and used spot instances for Gitlab nodepools and Jenkins build agents. In some instances we replaced ALBs with NLBs because it was cheaper based on that particular use case. There are many innovative ways to drive down costs but it takes time and experience.
Great video Daniel, cost is very important and you covered relevant opportunities to optimize your costs. I would like to add some other to your list: CloudWatch data ingest: Watch the amount of logs you put on your CloudWatch logs because you will be charged for data ingest, not only log retention period. DynamoDB: Stay way from scan operations, and keep an eye one Backups Plans or POITR Lambda: fine tune the memory, lower memory not always means lower lambda costs Thanks Daniel, great video 🎉
You were spot on. Great video and so happy that you captured the cost of data transfer. Many organization miss this and the cost can be hefty. The rest of the costs can be made predictable, especially if you're signing a multi-year agreement to reserve resources.
That last tip (taking cost in consideration right in the design phase) was so spot-on. THANK YOU !! I am working on a new small app and I have to look at the whole thing again. THANKS A LOT. The other one I find so helpful is your suggestion about those Cloudwatch log files. Have to do something with those. Once more, THANK YOU VERY, VERY MUCH !!!
Right sizing: Ex2: cpu utilisation memory and network utilisation. S3: General purpose, Ingres egres Data Retention Pricing calculator Design with cost in mind
Cloudwatch logs! I had a ton of debugging in my Lambda functions and when we went live the Cloudwatch costs went through the roof. Restrict how much you log (have a switch for different levels of logging) otherwise Cloudwatch will kill you.
I have a doubt at 5:40 - 6:00 time duration, AWS does not charge for data transfer within Availability Zones (AZs) of the same region. When data moves between resources located in different AZs within the same AWS region, it is considered to be "in-region" data transfer, and it is not subject to any additional data transfer fees. Would you please give clarity on this ?
The last organization I worked for we were actively trying to reduce the company's AWS costs over a one year period. A few things I've learned from that experience was that VPC endpoints do help to reduce costs. Although you have to pay for them but by utilizing VPC endpoints, we can reduce egress traffic costs to S3 and improve security by keeping your data transfers within the AWS network, while also benefiting from reduced latency since the traffic stays within the AWS infrastructure. We recommended that the company use reserved EC2 instances, got rid of EFS and used spot instances for Gitlab nodepools and Jenkins build agents. In some instances we replaced ALBs with NLBs because it was cheaper based on that particular use case. There are many innovative ways to drive down costs but it takes time and experience.
Smart
Great video Daniel, cost is very important and you covered relevant opportunities to optimize your costs. I would like to add some other to your list:
CloudWatch data ingest: Watch the amount of logs you put on your CloudWatch logs because you will be charged for data ingest, not only log retention period.
DynamoDB: Stay way from scan operations, and keep an eye one Backups Plans or POITR
Lambda: fine tune the memory, lower memory not always means lower lambda costs
Thanks Daniel, great video 🎉
Straight to the point, no draw outs. Thank you for making this.
You were spot on. Great video and so happy that you captured the cost of data transfer. Many organization miss this and the cost can be hefty. The rest of the costs can be made predictable, especially if you're signing a multi-year agreement to reserve resources.
Glad it was helpful!
That last tip (taking cost in consideration right in the design phase) was so spot-on. THANK YOU !! I am working on a new small app and I have to look at the whole thing again. THANKS A LOT. The other one I find so helpful is your suggestion about those Cloudwatch log files. Have to do something with those. Once more, THANK YOU VERY, VERY MUCH !!!
You're very welcome and I'm glad it was helpful!
Right sizing:
Ex2: cpu utilisation memory and network utilisation.
S3: General purpose,
Ingres egres
Data Retention
Pricing calculator
Design with cost in mind
My favorite aws channel!
Thank you so much!
Nice straight to the point explanation. Thank you 👍🏻
Cloudwatch logs! I had a ton of debugging in my Lambda functions and when we went live the Cloudwatch costs went through the roof. Restrict how much you log (have a switch for different levels of logging) otherwise Cloudwatch will kill you.
If you want to reduce costs do this:
use spot instances and autoscaling. Obviously for clients we do much more.
Great video Daniel. Super important topic
How come it has less views.. KILLER content.. Great job!!
Apparently, a lot of people took AWS Architect Associate exam 😉 I agree, great content!
I have a doubt at 5:40 - 6:00 time duration,
AWS does not charge for data transfer within Availability Zones (AZs) of the same region. When data moves between resources located in different AZs within the same AWS region, it is considered to be "in-region" data transfer, and it is not subject to any additional data transfer fees.
Would you please give clarity on this ?
Another cost optimization tip that I can think of is to shutdown the services that are not required during business hours.
Great one!
Aws egress is extortion
hui kiang sue kren 😂😂😂😂😂😂😂😂
Are you ai or real ?