Cheap but resilient: the Kubernetes HA lie
Overview
I keep seeing the same tension: everyone wants "cheap but resilient", nobody wants to pay for the second replica. That Reddit thread on zero downtime drains for single-replica workloads is ju
fieldnoteswithabhinav.hashnode.dev5 min read