Logistics: 24-hour cargo on always-on cloud

Cargo does not wait for a server to be rebooted

The problem

Logistics and terminal operations do not have an off-hours window. Freight keeps moving at two in the morning the same as at two in the afternoon, but the systems underneath were often sized and maintained for a quieter era, before volume or the uptime expectation reached what it is now. A hardware fault, an unplanned patch or a spike in load does not wait for a convenient time either, and when it lands the operation is either degraded or stopped while somebody works out what broke.

The usual answer

Operators compensate with people rather than infrastructure: a standing on-call rota paged at any hour, a manual failover procedure someone has to execute correctly under pressure, and a maintenance window scheduled for the quietest hour available, which in a 24-hour operation is still a real hour of exposure. It works, at the cost of the fatigue that comes with genuinely round-the-clock coverage and the risk that comes with a manual step in the middle of an incident.

How we approach it

We move the operation onto managed cloud infrastructure with monitoring, backup and recovery built into the platform rather than executed by hand. Failure detection and failover are automated where the workload allows it, patching happens on a schedule that does not require someone to time it around live traffic, and cost is monitored alongside uptime so resilience does not come with a runaway bill attached. Where a specific system cannot fail over automatically, that gets identified and planned around explicitly rather than discovered mid-incident.

What changes

A hardware failure becomes an alert someone has already been notified about and a system that is already recovering, rather than a phone call and a morning lost. The operation keeps running at the hours it actually runs, and the team on call responds to fewer incidents because more of them never reach a person.