Hacker News Reader

But not having swap will cause the machine to misbehave (https://utcc.utoronto.ca/~cks/space/blog/unix/NoSwapConseque...). Swap in this situation acts like a backup, surely you would prefer temporal degradation then total system shutdown. The solution here is continuous monitoring of free memory (we use datadog) + swap.

View on HN

> surely you would prefer temporal degradation then total system shutdown

Sometimes, but sometimes you want the opposite.

A slow service (e.g. a db server with a failing disk) might still pass health checks but degrade overall system performance.

It is easier/simpler to detect a total outage than a slow/flapping service.

Reply on news.ycombinator.com