Comparison
Maintenance windowvsPatch cadence
Maintenance window
the managed database will be restarted by the provider at some point inside a four-hour window you chose months ago.
A declared period in which disruptive operations — provider patching, node replacement, version upgrades — may occur. Managed services apply them whether or not you have planned for them, so the window is a commitment you have made even if nobody remembers making it. Treating provider maintenance as a routine failover you must survive, rather than as an event you must be awake for, is the difference between a mature platform and a fragile one.
Full entry →Patch cadence
you fell four versions behind and now the upgrade is a project, because you can only move one version at a time.
How regularly the platform's own components — nodes, orchestrator, plugins, base images — are updated. Falling behind compounds nonlinearly: support windows close, upgrades must be taken in sequence, and each step's incompatibilities arrive together. Frequent small upgrades are boring and cheap; infrequent large ones are a quarterly project with an incident attached.
Full entry →