Maritime Safety → DevOps Lessons
Yesterday's dramatic oil tanker collision off the Yorkshire coast highlights critical parallels between maritime safety and DevOps practices. Here's what we can learn:
The Stena Immaculate was anchored overnight, awaiting a berth to be able to unload. There were several other large ships in the same area doing the same thing, but all outside of the charted anchoring area. Documentation drift leads to compliance and safety issues
The Solong cargo ship was steaming down the east coast of the UK following the same route that the crew had done many times. The crew will have assumed that this journey would be exactly the same as the previous ones. Be aware of complacency and test your assumptions
Because the Solong cargo ship was sailing through a busy area, the crew will most likely have reduced the sensitivity of alarms. Whilst that is perfectly fine to do, it requires maintaining a good lookout. If you turn off alarms during some of your processes, monitor the situation using other methods
After the collision, distress calls were immediately made, a well-rehearsed process kicked in, and nearby teams immediately responded to the call. Make sure your incident management processes are documented and tested, and be ready to support other teams without getting in the way. Well-documented processes enable swift action
It will be many months before we hear from the Marine Accident Investigation Board about exactly what happened here, but they will have been informed as soon as possible by either the captains or the most senior survivor. Make it easy for engineers to report incidents through a simple form and have processes in place to both review incidents and disseminate the findings so that other teams can learn from them
- What similar analogies have you discovered between other industries and DevOps?
- How do you currently handle system monitoring when certain alerts must be disabled?
- What incident response protocols could be strengthened in your current workflow?