Deployments
The Deployments page (sidebar → Deployments) lists all the clones you've ever run, and it lets you drill down into the details of any of them, either live or after they're over — even a graph in real-time data per node and a console for each machine involved.
The list of deployments
A filter bar allows you to narrow down the list using free-text search or status (All / In progress / Completed / Error / Stopped), prescription, or whether the deployment used a saved image in front of the live master disc, plus an order (more recent, faster, longer, or with more nodes). On top of that, a few Summary cards show totals for ongoing and completed deployments and with an error, and an average client speed based on what the current filter.
Columns
| Column | What's on display |
|---|---|
| ID | Short identifier for the deployment. |
| Status | In Progress / Completed / Error / Stopped. |
| Recipe | The recipe used. |
| Nodes | How many nodes are involved? |
| Customer Service | Current transfer speed for clients. |
| Master's Degree | Current transfer rate in the master's program. |
| Duration | Elapsed time (a min-max range once the nodes begin to complete at different times). |
| P2P | Average proportion of data that each client received from other peers rather than from the master. |
| Progress | Overall completion bar. |
| Started | When did the deployment begin? |
Clicking anywhere in a row displays an inline panel with KPI cards (nodes, average/minimum/maximum client speed, speed master's program GPA, duration, P2P average), any conflict at the level of host detected (as a mismatched disk), the complete set of recipe parameters as they were at the time of deployment, and one table per node (role, status, progress, speed, P2P ratio, duration, chunks received, disk health, network interface).
Row Actions
- View opens the full panel, described below.
- Download diagnosis (displayed only when the prescription was (configured to save telemetry) downloads the raw telemetry via A tick that recorded netdd during execution.
- Stop (while it's running) asks if you just want to stop it, or stop it and delete it directly.
- Retry (after completion or failure) triggers a deployment Completely new, using the same recipe, destination nodes, master, and image — useful for rerunning a failed batch without reconfiguring it Nothing.
- Delete (once it is no longer in progress) permanently deletes it the deployment history.
Select multiple rows to delete them in bulk — a display in A course that is caught up in a mass selection is automatically paused before to be eliminated.
Deployment States
| Status | Meaning | What You Can Do |
|---|---|---|
| In progress | In progress. | Stop him. |
| Completed | Successfully completed. | Repeat, delete, download diagnostics. |
| Error | It ended with a ruling. | Repeat, delete, download diagnostics. |
| Arrested | Cancelled by an operator. | Remove it, download the diagnostic report. |
The Deployment Panel
Opening a deployment displays a live panel (or, once it's finished, a final panel): real-time charts if the process is still in progress; otherwise, the same frozen charts back in its place once it's finished.
The header displays the name of the recipe/group, the source of the master, when it started, a Source bar (what percentage of the bytes came from the master versus peer-to-peer redistribution), and the speed of Combined transfer of all clients. A Consoles button opens a live terminal for each node in the deployment (see below); a button The Telemetry download message appears if the recipe was configured to Record it; Stop appears while the display is still in progress.
Row of KPIs
| Card | Meaning |
|---|---|
| Active Customers | How many client nodes are currently reporting? |
| Completed | How many customers have finished. |
| Total chunks | The transfer size, in chunks. |
| Active time | How long has the deployment been underway? (It freezes when it's finished.) |
| Master's Program Duration | The current graduation rate for the master's program. |
| Master's program timeout | Countdown until the recipe's "stop if no client connects" timeout (∞ if disabled). |
| Average customer speed | Average transfer speed among all reporting customers. |
Statistics Table by Node
One row per node (with the master listed first): role (master/client), what What percentage of this client's data came from the master's program compared to other sources? peers, current/minimum/maximum/average speed, and an ETA — minutes remaining while it is in progress, or the total time spent once it is completed.
Customer Progress Chart
One row per customer, showing the percentage completed over time, with a selectable time range (from 1 minute to the entire deployment) — Useful for seeing whether customers are clustering together or spreading out.
Link Health
Each customer receives a health score out of 100, based on three indicators: speed relative to the rest of the group, proportion P2P** (peers are better than relying entirely on the master server) and stability (how consistent that customer's speed has been (recently). The scores are color-coded—green for the nominal, blue for normal, amber for degraded, red for the critical — and each row includes a diagnosis in plain language (e.g., "no "P2P," "overloaded master," "unstable network"), with a new client Displaying "warming up" until it has enough samples to calculate a score.
Additional details appear below a customer's rating when applicable:
- Multicast reception — packet loss, how many were recovered by correcting errors going forward rather than asking relay, how many blocks relied entirely on peer-to-peer, and any overflow of the receive buffer — is displayed in deployments that use multicast transport.
- Receive Speed Limit — current speed versus the the client's configured limit, and how many seconds it has been limited.
- Differential Cloning Savings — how many blocks were skipped because the change made, and the resulting percentage of traffic saved, for deployments that use incremental diff.
Real-Time Speed Graph
One line per node, transfer rate in MB/s along the time, using the same range selector as the progress chart, and a switch between Download (received data) and Upload (P2P) (data (that a client is redistributing to its peers).
Metrics Chart
A second graph lets you plot other counters per node in the same way over time: network errors (dropped packets or packets with CRC errors), TCP transmissions and disk temperature, with a rotation mode that automatically switches between nodes or stays on the one that seems most problematic. If the prescription triggered the more in-depth diagnosis, here Additional data is also displayed—link speed, round-trip time and TCP round-trip time, disk read/write speed and latency in percentiles, tail depth, CPU, and RAM.
Live Console
The Consoles button opens a grid of live terminals — a one card per node in the deployment, each representing a real, interactive console over a secure WebSocket connection. The cards can be dragged, resize, minimize, or maximize, and a sidebar lets you search and Choose which nodes to display. Only one person can have the console open at a time. from a given node at a time, and the consoles are only accessible while the The deployment of that node is still in progress.