I have the same outlook about serverless. Moving things out of process means that you now have to deal with new failure modes and latencies for communication between components. It works exactly the same as when a company grows and splits into departments. Only it doesn't make sense to split into 5 departments (or any at all) when your company only has 3 people. I think this is a lesson from Conway's law that has been missed as people are discovering serverless.
On AWS, there is an interesting ops middle ground between shepherding individual EC2 machines to using Lambdas. And that is using Fargate. Unfortunately it is more expensive than both of these options. We are using the similar but cheaper option -- our own ECS cluster. It is like fargate, but we have to tell it what EC2 resources should back the cluster. They only charge for the EC2 resources -- the clustering is no charge.
I run technology at Ingenious, an agency specializing in building products for the healthcare industry. 💛 remote work ‧ sometimes 🔨 are the best tool for the job.
Nice info about how to handle it on AWS. I'm currently using Hatchbox.io to provision DO boxes, both in cluster or standalone, also I tend to over use Heroku for dev/staging/review apps which provisions a new Dyno and DB for every new PR.
For further actions, you may consider blocking this person and/or reporting abuse
We're a place where coders share, stay up-to-date and grow their careers.
I have the same outlook about serverless. Moving things out of process means that you now have to deal with new failure modes and latencies for communication between components. It works exactly the same as when a company grows and splits into departments. Only it doesn't make sense to split into 5 departments (or any at all) when your company only has 3 people. I think this is a lesson from Conway's law that has been missed as people are discovering serverless.
On AWS, there is an interesting ops middle ground between shepherding individual EC2 machines to using Lambdas. And that is using Fargate. Unfortunately it is more expensive than both of these options. We are using the similar but cheaper option -- our own ECS cluster. It is like fargate, but we have to tell it what EC2 resources should back the cluster. They only charge for the EC2 resources -- the clustering is no charge.
Nice info about how to handle it on AWS. I'm currently using Hatchbox.io to provision DO boxes, both in cluster or standalone, also I tend to over use Heroku for dev/staging/review apps which provisions a new Dyno and DB for every new PR.