Pending Inputs
The classification backlog: stops captured by telemetry but not yet classified. It's the origin of the "Pending" downtime seen elsewhere, and the single biggest driver of data quality.
1 · The purpose of this report
It turns "our data has gaps" into a worklist: exactly which operators, products and reasons need classifying, and how many hours hang on it. Clearing this queue is the prerequisite for trusting every other report, it shrinks the Pending and Unknown-operator distortion everywhere downstream.
2 · On the page
Top to bottom: the controls, then the summary, then the worklists by operator, product and reason.
2.1 · Selectors
| Control | What it does |
|---|---|
| Point selector | Choose the site or line whose classification backlog you're clearing. |
| Date selector | Set the period. Review it daily at shift level so the queue trends to zero rather than accumulating. |
2.2 · The summary
The pending count, the pending time, and the share of paid time sitting unexplained until it's classified.
2.3 · The operator worklist
Who has classification debt, ranked, so you can see whose stops aren't being coded, and, usually at the top, the "Unknown operator" bucket where nobody was signed on at all.
2.4 · The product and reason worklists
The same queue cut by product and by stop reason, so you clear the largest contributors first rather than chipping at the count.
3 · Use cases
- Do this first. Clearing the queue is the prerequisite for trusting the Executive Summary and Performance Overview, a large Pending bucket there traces straight back to this page, so it's the first move in any improvement order.
- Fix capture, not just classify. The largest single bucket is usually "Unknown operator", a sign-on gap to fix at the point of capture, so events attribute themselves in future.
- Track deployment health. A healthy site trends this queue to zero, not up, it's the leading indicator that a rollout is working.
4 · Good to know
Work the queue top-down by hours, not by count, and review it daily at shift level. Augos doesn't hide unexplained time, it measures it, prices it, and hands you a named list to clear it. That honesty is the point: the "Pending" you see elsewhere is real, measured loss awaiting a reason, not a gap swept under the rug.
5 · Common questions
Why does "Pending" dominate our Executive Summary?
Because this queue is unworked. It's real, measured loss awaiting a reason, a classification task, not a report defect.
An operator shows as "Unknown operator".
The stop wasn't signed to anyone at capture. Fix operator sign-on so events attribute automatically, it's usually the single largest bucket here, and fixing capture beats classifying after the fact.