Best-effort high availability for single-replica Kubernetes workloads: no downtime from announced node disruptions (drains, upgrades, consolidation, spot reclaims) without a standing second replica.
A Python framework to write Kubernetes operators in just a few lines of code