Troubleshooting an issue like this when I have the time and am prepared for a potential outage (with human resources hot and standing by for immediate action) is VASTLY different than encountering it the evening before some critical deadline for a multi-million dollar project (as Murphy's Law will be sure to have it).
When I have control over the deployment of updates I can push them through my own QA environment first, to uncover many of these kinds of issues before they hit production. Vendors pushing them out on their whim leaves me subject to whatever fast and loose practice they use and prevents me from being able to properly manage my own infrastructure.
A slow rollout certainly helps but doesn't satisfy the kind of 9's I demand in the environments I care for.
Troubleshooting an issue like this when I have the time and am prepared for a potential outage (with human resources hot and standing by for immediate action) is VASTLY different than encountering it the evening before some critical deadline for a multi-million dollar project (as Murphy's Law will be sure to have it).
When I have control over the deployment of updates I can push them through my own QA environment first, to uncover many of these kinds of issues before they hit production. Vendors pushing them out on their whim leaves me subject to whatever fast and loose practice they use and prevents me from being able to properly manage my own infrastructure.
A slow rollout certainly helps but doesn't satisfy the kind of 9's I demand in the environments I care for.