Code and history
Adding runner token usage to observability
I want to add details about token usage on the runners to the observability system. What subsystems are affected? When was this last modified, and why?
Two bounded factory investigations
Questions followed through the graph.
Each walkthrough begins with a small coding question, shows the Cypher and reviewed captured result, then explains what the evidence can and cannot support.
01 · CODING AGENTasks
02 · GRAPH AGENTplans
03 · CYPHERretrieves
04 · GRAPH AGENTchecks
05 · CODING AGENTacts
Prior research
When an earlier investigation changes a diagnosis
The monitor says a runner is idle. Does that mean it is dead, and what should I check before I plan a repair?
How to read this gallery
The public site contains a reviewed projection, not a source database. “Captured result” means the displayed Cypher ran against an isolated, read-only demonstration graph and the selected rows were checked against a pinned capture before publication. Live NeoDash and source access remain protected.
Gallery implementation revision: 3e800bb13f21683a7bc34d4f443acf04422d90a0. Read the evidence and privacy boundary.