Hacker News Reader

Comment by grey-area | Hacker News Reader

It's pretty easy to cut down clusters to the minimum number of nodes for them to function (say cutting 10 servers to 2), or downgrade the size of instances to compensate for lower traffic. There's no need to restructure infrastructure in most cases, if anything using a cloud provider makes this easier.

View on HN

It's "easy" to do this, but it's also just as "easy" for your service to fall over when you do it wrong. Capacity planning and performance engineering is a lost art in a world where you would simply throw money at the problem until it went away.

It's not that hard (I've done this recently for clients). If your traffic is 90% of normal for a few months, scaling down with load makes a lot of sense, then you can scale back up when ready. There are also autoscaling solutions like serverless.

This is true for simple services. It's different when you're dealing with something like a Cassandra or Elasticsearch cluster, where the system isn't stateless and scaling up takes some time for re-balancing.

Sure, it won't apply to all services, particularly datastores are best left alone, but if you need a quite few web workers say which sit in front of your search or db to service millions of users, and can cut that in half during low traffic times, you may save significant money, as the user pram above has noted.

Aren't you making gp's point for him?

GP was saying there is no flexibility of infrastructure if you use cloud servers and they are therefore recession proof.

I was saying it's pretty flexible within the constraints set up, and often companies have spare capacity they could significantly scale down to save money (for example a travel company during a pandemic has very few customers), so it is flexible in the most important way (cost).

That means cloud companies will see a significant reduction in resource usage and thus revenue if enough customers do this.

Reply on news.ycombinator.com