SMART INSIDER / PROCESSING

Laptop Observatory

Collection, calculation and checking — before bulk publication.

Awaiting telemetryNo snapshot received
Preparing the handoff

The original database is being preserved while the laptop processing system is built.

Six-minute output watch

Waiting for a fresh full-window measurement.

Counts full or partial portfolios durably saved by a worker or the laptop master. Independent verification and website publication are separate stages. No-history assessments and retries do not count.

Year coverage

Actual laptop calendar sweep

The total colored bar length is the portion of weekdays actually swept by the laptop; weekends are excluded and untouched dates remain empty. Within that processed portion, dark green is fully verified, pale green is verified no-prior-history, yellow is partial, red needs review, and orange waits for input. Public publication is separate; 2026 ends at the source cutoff.

PRIMARY THROUGHPUT MEASURE

History lines examined One line = one prior canonical history record examined for a portfolio during input preparation. Counts appear as that examination finishes, even if prices or later checks are still pending. Repeating the same portfolio and history snapshot adds nothing; examining changed history is new work. These are input activity counts, not completed portfolios, publications or a hardware benchmark. Large histories that later need review can create high scan records. Checked calculation lines remain separate below.

This input-examination counter starts with its own measurement history. Records use full rolling windows and survive restarts; an explicit reset starts a new comparison period while preserving the old records. Each date marks the end of the record window, in your local time. Partial windows and test runs cannot set records.

Awaiting processing telemetry

Laptop resource usage

CPU: caution at 80% · ease back at 90%
Total CPU in use—
RAM in use—

Resources reserved for coordination, verification and other applications.

Latest scaling decisionWaiting for production measurements

Portfolio throughput Counts unique portfolios when independent checks first pass in each category in this run. A partial portfolio later completed can appear once in each category; the combined total counts category milestones, not distinct portfolios. Each window shows full / partial = total. No-prior-history assessments are separate; retries and repeated checks add nothing. Portfolio sizes vary, so these counts complement history lines. These are local results awaiting later publication, not published completions. Early windows show only the history observed so far. Separate full, partial and combined-total best-since-start records use full rolling windows, survive restarts, and show the window end in your local time.

Locally verified cash portfolios

Throughput by stage

Counts · hourly rates underneath

Awaiting preparation telemetry

Stage / work unitWaitingLast 6 minLast 1 hourLast 10 hoursLast 20 hoursLast 200 hours

Read from history examination above, through input resolution, checking and master import. The units differ between stages and must not be added together. Shared jobs may unblock many portfolios. Partial windows use the observed duration; no estimated historical counts are backfilled.

What is holding portfolios?

Locally verified this run—Cash and no-history shown separately below
Verified / last hour—Unique accepted cases · observed rate
Calculation workers—Adaptive concurrency
Waiting for checking—Calculation results awaiting acceptance

Machines & provider activity

Awaiting provider telemetry
MachineRole / measureActivityLast 6 minLast 1 hourRecorded total

Yahoo and SEC rows count actual responses, with HTTP 200, not-found (404), and other responses separated. SEC cache reuse is excluded from external fetches. These counts are not added to portfolio history lines or verified completions. Virginia activity comes from its request receipts; this is not a remote CPU monitor.

Line throughput over time

Last hour · history lines per minute

Processing history will appear after activation.

Throughput by current assignment

Lines examined · hourly rates underneath
WorkerLast 6 minLast 1 hourLast 10 hoursLast 20 hoursLast 200 hours

Current assignments appear above. Worker numbers identify assignments over time, not additional laptops. Finished assignments consume no worker CPU. Early rates use the observed portion of the window.

Finished assignments · historical records

These are retained records, not running or idle processes available for reassignment.

Finished assignmentLast 6 minLast 1 hourLast 10 hoursLast 20 hoursLast 200 hours

Processing stages

Input work and checking remain separate from completed results

Workers

Independent assignments and output stores
Worker / machineAssignmentStageCases assessedLocally verifiedLast signal

Outcome accounting

Migration & publication

The Mini’s original databases stay frozen. Public publication is a later, separately verified step.

Recent worker adjustments

Compare against the throughput windows above