Patching without a maintenance window
The traditional answer to updates is a scheduled outage. Forecourts and highway sites do not have one to offer. What works instead is batching: a few machines at a time, during each site's quietest hour, verified before the next batch starts.
Updating without interrupting service
A badly planned update can take several terminals in the same branch out of service at once, right at the busiest time of day. That is why rollout happens in batches, inside an agreed window, with post-checks and the option to roll back. The aim is for the fleet to end the day with the same availability it started with. It is one of the questions branches ask us most often: terminal out of service.
A bad release should not be a crisis
Every release carries some risk. What determines whether that risk becomes an incident is how quickly the previous version can be restored, and how many machines were exposed before anyone noticed. Both are decided by how the rollout is structured.