The observability gap in every company
Modern engineering teams would never run a production system without metrics, logs and traces. Yet the same companies run their operations, where most payroll is spent, with no instrumentation at all. Management sees outcomes and headcount, and nothing in between.
The gap has a cost. A process nobody can observe cannot be improved systematically. Improvement then depends on who happens to notice a problem and whether they think to mention it, which is intuition, not engineering.
What workflow observability measures
- Time distribution: which processes and application categories consume the day
- Repetition: sequences performed identically, many times, by many people
- Friction: tool switching and waiting
- Drift: how the real process differs from the documented one
- Change: whether an automation or redesign actually moved the numbers
Observability without surveillance
Server observability never asks which engineer caused CPU load, and workflow observability should never ask which employee is slow. The discipline works at the process level: data is minimized at the source, identifiers are pseudonymized, and insights are aggregated. Nodra implements this as architecture, not policy: the data needed to judge individuals is never collected.
Treated this way, observability is pro-worker: it finds the work the system forces people to do unnecessarily and removes it, instead of asking anyone to work harder.