It just a time limit of the life of a single MicroVM.
replyUsing this for a long lived "developer environment" would be extraordinarily expensive anyhow. Scaling the vCPU + RAM cost of these to the same shape compute optimized Graviton On-Demand EC2 instance (16 vCPU x 32 GB RAM) shows about 4x the cost.
So don't do that. Just use an EC2 instance.
But these have near instant suspended/resume, and they even have vertical scaling of the ram, which is a great feature that’s not very common.
replyIn theory, you could set up a process to move data/filesystem between sessions into and out of s3.
replyThey are long-lived if you're a mayfly.
replyBut I think the point is that they should be cheap to set up, and because of the short life, never really contain anything except the potential to compute when needed, not important data.
You can use them for dev environments.
replyYou just have to finish development in 8 hours.
I'm assuming you can launch them again after 8 hours.
replylambdas are ephemeral on compute, but couldn't you connect up EFS for your long lived data?
replythen when you launch the next one, its like you are still there?
EFS is extremely slow for many workloads. We tried it for builds and various other common use cases for coding agents and the performance just isn't there. I'm guessing lots of small random reads/writes just isn't going to ever work well.
reply