Remote Collection

Retrace can collect terminal activity from remote hosts over SSH — without installing anything on the remote machine.

How it works

The remote only needs an SSH daemon and a shell. All parsing and redaction happens locally at ingest; nothing is installed remotely.

graph LR
    A[Local Retrace] -->|SSH| B[Remote host]
    B -->|history + spool data| A
    A --> C[Redaction engine]
    C --> D[(Local DB)]

Registering a remote

From the Settings → Remotes tab (or retrace/remote.py):

Field	Example
Name	prod-web-01
Host	192.168.1.50
User	deploy
Port	22

Then click Test to verify the SSH connection, and Collect (or run retrace collect-remote) to pull data.

Requirements on the remote

• SSH daemon running
• A shell (bash/sh) accessible via SSH
• Read access to the user's shell history

No agent binary, no root, no installation.

Security notes

• Remote collection is opt-in and explicit — nothing connects outbound automatically.
• Data is redacted locally at ingest, same as local sources.
• Credentials for remotes are stored in your local config — keep the config file permissions locked.

Managing remotes from the UI

• Remotes tab — see registered hosts + per-host record counts
• Settings → Remotes — add / remove / test hosts
• Settings → Actions → collect remotes — pull from all registered hosts in a background job (pollable via the jobs API)