Aerie Overview

Aerie is the live agent-operations workboard for Reasoning Flows. It gives you a real-time operational view of every project and object in your workarea—so you can watch executions as they happen, spot failures the moment they occur, and act without digging through individual project logs.

Where Reasoning Flows is where you build and Reasoning Atlas is where you understand structure, Aerie is where you operate.

---

Accessing Aerie

Aerie is available from the sidebar under Reasoning Flows → Aerie, or from the Aerie card on the AI Orchestrator home.

The board opens in a live, continuously synced state—the REASONING OPERATIONS • LIVE indicator in the top bar confirms real-time sync and shows when data was last refreshed.

---

The Workboard at a Glance

Aerie is organized in three layers, from broadest to most specific:

---

Global Health Strip

Five live counters at the top of the board summarize the operational state of the entire workarea:

Each counter is color-coded—green for running, blue for completed, amber for action required, red for errors—so a single glance tells you whether the workarea needs attention.

---

Project Columns

Each project in the workarea appears as a column on the board, identified by its workspace code (for example, WP-92). Columns display per-project operational KPIs:

Below the KPIs, a row of status dots summarizes how many of the project's objects are in each state, giving you a per-project miniature of the global health strip.

Each column also includes an external-link button that opens the project directly—Aerie observes; changes to flows and objects are made in the project itself.

---

Object Cards

Inside each column, every object appears as a card showing:

• Object identifier (OBJ-XXXX) and runtime badge (for example python-3.9, php-7, dk-data-pipe-mysql)
• Object name
• Type tag — ETL, Transform, ML, LLM, or Render
• Live status — Running (green), Done, or Error (red)
• Average duration of recent runs

When an object fails, the error trace appears directly on the card with its timestamp—you see what failed, when, and the first lines of why without leaving the board.

Tip: Cards with the Error status are your fastest entry point for incident response: read the inline trace, then use the column's external-link button to open the project and fix the failing object.

---

Searching and Filtering

The top bar provides two ways to narrow the board:

• Search — Find projects and objects by name across the entire workarea.
• Type filter — Show only columns of a given Project Type: Batch, Interact, API, MCP, or WARP.

Tip: Use the type filter to isolate a single operational surface—for example, show only WARP services when checking the health of persistent endpoints, or only Batch projects during a scheduled-pipeline window.

---

Common Use Cases

• Morning operational check — Open Aerie, read the health strip, and clear any Errors or Action Required items before the day's runs.
• Incident response — When an alert fires, Aerie shows the failing object and its error trace immediately, with a one-click path to the project.
• Cost and performance awareness — Tok/Hr and Avg Run per project surface expensive or slow workloads at a glance.
• Deployment watch — After releasing a new flow, keep Aerie open to confirm the first runs complete successfully.

---

Who Uses Aerie

Aerie serves everyone accountable for runtime, with each role reading a different layer of the board:

Note: Aerie shows that something failed and the start of why. Remediation happens in the project itself—use the column's external-link button to jump from observation to action.

---

How Aerie Fits the Reasoning Flows Cluster

---

Summary

Aerie turns the question "is anything broken, where, and since when?" into a single screen. With a live health strip, per-project KPIs, and object-level status and error traces, it provides the operational visibility layer for everything built in Reasoning Flows.