Hacker News new | ask | show | jobs
by kevmo314 495 days ago
I once had a project that moved a microservice into a monolith on a larger instance type. SRE flagged that larger instance type as expensive, at $1k/mo. They really drilled in, asking "are you sure? that's a very expensive instance size". Yes, I was sure. It was replacing a service that was costing $20k/mo across a number of tiny instances.

Now I have a strong belief that paying people whose only job is to save on costs is a bad move: they will conclude that the best way to save on costs is to shut down the company. Technically correct!
2 comments
selix 495 days ago
That's a perfect example of the type of thing I was trying to get across at in the article.

Cost optimization + a lack of systems thinking = frupidity.

link
daxfohl 495 days ago
Similar story, forced to downsize a service, but it had one RPC with high availability requirements. So we had to spin up another service for this one RPC. And a pipeline to copy the data. And a duplicate database. And now the high availability data is stale, requiring other teams to do weird workarounds that also increase cost and make things more buggy. Yay, frugal.
link