Comparison
Bus factorvsSingle point of failure
Bus factor
one person was on holiday and the release could not go out until they got back.
The number of people who would have to be unavailable before work stops. It is usually one, and it is usually one for reasons nobody chose: the person who set up the pipeline, the only one with the production credential, the only one who understands the pricing rules. It is worth raising in planning rather than after the fact, because the fix — pairing, documentation, rotating who does the scary task — takes weeks and only works before the absence.
Full entry →Single point of failure
you trace the dependency graph and find one component whose failure takes everything else down with it.
A component with no redundancy whose loss causes total failure. They are easy to spot in the data plane and easy to miss in the control plane: the config service, the DNS zone, the single deploy pipeline, the one person with production access. Redundancy that shares a failure mode is not redundancy.
Full entry →