Observe: Knowing what the system is doing

When something goes wrong in your review pipeline, it's often because you didn't have a clear picture of what the system was doing before the change.

Telemetry as Context

Agents write code based on the context they are given. If you don't have good telemetry in production, you can't provide that context.

// Example of logging that provides good context
logger.info("Processing user request", {
  userId: user.id,
  action: "merge_pr",
  durationMs: 120,
});

Observability is the first stage of the loop for a reason. You can't fix what you can't see.