Non-interactive by design
There is no terminal session and no prompt. Each command runs to completion exactly as typed, and its output, exit code and duration are recorded. A command that waits for input receives none and exits.
That is what makes it practical to run a command across many servers, and what makes the record of it reliable.
Choose servers by name, group, tag or OS
Target servers by name, group, tag or operating system, or select every server on the account. Before anything runs, your selection is resolved into a list of servers, with known problems flagged: never connected, unreachable at the last check, or unable to become root.
The list is only a preview. No server is contacted until you start the run, and each server is checked again when it does.
Live output from every server
Each server has its own status and log, updated while the command runs. Open any server to read its output, or download the log as a text file.
Every command has a time limit between one minute and one hour. A command still running at the limit is stopped on the server itself, so it cannot keep running after you close the page.
Retry only the servers that failed
If a command fails on three of forty servers, a retry starts a new run on those three, along with any servers that were blocked, skipped or cancelled. The new run links back to the original, and servers that succeeded are not touched again.
Cancelling stops the command on servers where it is still running and skips servers that have not started. Both actions are recorded on the run.
Run as root or as the connecting account
Run a command as root, directly or through sudo, or as the account the platform connects with, for servers where that account does not need root.
Commands and their output are kept in the run history, where everyone on your account can read them. Do not include passwords, tokens or keys in a command.
Frequently asked questions
- Is this a web-based SSH terminal?
- No. There is no interactive session. Each command runs to completion on every selected server, and the result is recorded as a run.
- Can I run a command only on certain operating systems?
- Yes. Select servers by operating system, for example every Debian and Ubuntu server, or by the groups and tags you already use.
- What happens if a server is unreachable?
- The preflight marks it as blocked with the reason, nothing runs on it, and the other servers continue. Connection failures are retried a few times before a server is marked as blocked.
- Can I stop a command that is taking too long?
- Yes. Cancel stops it on servers where it is still running. Every command also has a time limit between one minute and one hour, enforced on the server.
- Who can see the commands I run?
- Everyone on your account can read the run history, including each command and its output. Commands are never shared outside your account.
Key points
- Target by name, group, tag, OS or all servers
- Live status and output for each server
- Retry failed servers as a linked run
- Cancel at any time; time limits enforced on the server
- Every command kept in the run history