As of April 2020, Fargate now supports EFS. aws.amazon.com/about-aws/whats-new/2020/04/amazon-ecs-aws-fargate-support-amazon-efs-filesystems-generally-available/
Excellent presentation, great points about security! Only proxy services or servers should be running in a public-facing subnet to minimize attack surface area from the Internet.
I was hoping to learn how NLB works in conjunction with multiple private subnets and how to make Enpointservice available in multiple AZs without running out of CIDR blocks. Does any one have any references?
12:10 Many containers may want way more than 4vCPU and 30GB ram. I was running Cassandra in Kubernetes and it would take 8 vcpu and 64GB ram (could have been more too).
I don't think anyone was suggesting that it'd allow you to do more computations - the goal of this is to reduce the overall complexity of operating large groups of pieces of software. For example, I've got a service that only needs to be running something like 10% of the time, and it's impossible to predict when it needs to be running beforehand. I only want it to run when needed, so I can define a fargate configuration for it and scale it up and down when needed without having to worry about additional logic & configuration that would need to be created and managed if I were to run this through traditional EC2 VMs
20:16 that is not correct, Fargate still does not support EFS mounts
As of April 2020, Fargate now supports EFS. aws.amazon.com/about-aws/whats-new/2020/04/amazon-ecs-aws-fargate-support-amazon-efs-filesystems-generally-available/
Excellent presentation, great points about security! Only proxy services or servers should be running in a public-facing subnet to minimize attack surface area from the Internet.
Nice presentation. Better if pricing comparison between ecs-ec2 / ecs-fargate for same service type and duration would be helpful. Any references?
I was hoping to learn how NLB works in conjunction with multiple private subnets and how to make Enpointservice available in multiple AZs without running out of CIDR blocks. Does any one have any references?
12:10 Many containers may want way more than 4vCPU and 30GB ram. I was running Cassandra in Kubernetes and it would take 8 vcpu and 64GB ram (could have been more too).
Tried to find the CF templates used but no luck.. Can anyone point to the resources he used in this presentation?
Found! gitlab.com/ric_harvey/bl_practical_fargate
Great presentation, thanks!
Nice presentation. Are the couldformation files available somewhere? Thanks
How can we know that this kind of summits, webinars or sessions are given by AWS?
What kind of terminal is he using? Is it iterm2 with oh-my-zsh?
Yep, with the Powerlevel9k/Powerlevel10k theme I think github.com/Powerlevel9k/powerlevel9k
I found this very hard to follow
can't ssh into your instance.. uh.. honestly it's a MASSIVE drawback - you can have ssh available and still do things properly.
You can, its just very cumbersome.
How come you can run more computations by running this docker. If you have 1 gb or memory it's 1 gb. This is just over complicated crap.
I don't think anyone was suggesting that it'd allow you to do more computations - the goal of this is to reduce the overall complexity of operating large groups of pieces of software. For example, I've got a service that only needs to be running something like 10% of the time, and it's impossible to predict when it needs to be running beforehand. I only want it to run when needed, so I can define a fargate configuration for it and scale it up and down when needed without having to worry about additional logic & configuration that would need to be created and managed if I were to run this through traditional EC2 VMs
Very slow and hard to understand
After this presentation amazon fired him.
Nope, he continued to work at AWS until Dec 2019