The live job console

Some jobs run for a long time on a customer's machine and do many things along the way. A golden image build, for example, runs its scripts on the master image or a clone: sealing it, optimising it, installing apps and running the image's own scripts. When something goes wrong half way, "failed" is not enough: you want to see what was run and what it said.

The Console on such a job's page shows exactly that. Every command the agent runs and its output show there as it happens, in order, and stay with the job afterwards. Nobody can type into it: it is a record, not a terminal.

Where you find it#

The console sits on the page of each run of a golden image. Open Infrastructure > Virtual desktops, the Images tab, then the image and the run. The console is open by default while the run is building, testing or rolling out, and can be collapsed. For a finished run it holds the full log.

When a job has nothing to show yet, the console says Waiting for the agent. A job that never sent a command to the device says This job sent no commands to the device.

What each line shows#

For every command:

• the time it ran (to the millisecond), the machine it ran on and the step of the job it belongs to
• the command line, with its parameters
• every line of output as it arrives, with error output marked
• the exit status when it finishes

Commands appear in the order they ran, with no gaps and no duplicates, even if your connection drops: the page fills in whatever it missed when it reconnects. A badge says Live while it is connected and Reconnecting while it is not.

Secrets are masked#

Passwords, keys, tokens and other secrets never reach the console. The agent masks them before a line leaves the machine, both the values the job knows (credentials it was given, a backup's encryption key) and anything shaped like a secret (a -password parameter, a JSON field named like a password or token, an Authorization: header, a private key block). The Tenvara server masks the same shapes again as the lines come in. A masked value shows as ********. The contents of files the job reads are shown only by their size.

Following, searching and saving the log#

1. While the job runs, the console follows the newest line. Click Pause to stop it scrolling while you read, and Follow to carry on.
2. Type in Search the log to filter the lines and highlight the matches.
3. Click Copy the log to put the whole log on your clipboard, for example to paste into a ticket or a message to a vendor.
4. Click Download log to save it as a text file named after the job.

Who can see it#

Anyone who can see the job can see its console: the module it belongs to is switched on, you have its view permission, and the customer is inside your scope. The console needs no extra permission and gives none: there is no way to write to it.

Kept with the job and its ticket#

The lines stay with the job, so its page always shows the full log. When a job that has a ticket finishes, its whole log is also attached to the ticket as a text file, so whoever picks the ticket up has it without opening the job.

How long console logs are kept is set in Settings > Data retention, under Job console logs, like your other logs. A log is never removed while its job is still running.

Tip: When a build step fails, search the log for the step's name or for "error", then copy the few lines around it into the ticket. The full log is already attached when the job finishes.

Related guides#

• Golden images
• Virtual desktops
• Audit log, retention and privacy requests