Many developers choose the cheapest cloud server configurations, like 2 vCPUs and 4 GB RAM, to run microservices such as Nacos. However, memory exhaustion can cause services to crash, SSH connections to fail, and processes to hang for over 200 seconds. This article argues that swap memory is not obsolete but a necessary safety net for such environments. It provides practical advice on configuring swap to prevent out-of-memory (OOM) kills, especially in test environments where stability is crucial. The post highlights that swap can be a cost-effective reliability hack, allowing developers to avoid upgrading to more expensive instances. For overseas developers managing budget cloud deployments, this is a timeless reminder that swap remains relevant for preventing unexpected downtime.
Swap memory is a critical safety net for small cloud instances running multiple services, preventing crashes and SSH lockups.