Skip to main content

Show which private worker executed a run (in run details + audit/log export)

We use private workers only, and one of the gaps for us right now is not being able to tell where a specific run actually executed once it finishes. While a run is active, we can sometimes spot worker IDs in the worker pool view, but after the run completes there’s no reliable way to trace it back to a particular worker or host. That becomes a real problem when users report repeated failures and we suspect one of our worker machines is unhealthy, rebooting, or having network issues. Without run to worker traceability, troubleshooting turns into guesswork.

What we need is simple. For every run, Spacelift could record and display which worker picked it up, and ideally which worker pool it came from. If there’s safe metadata available like a worker tag, label, or hostname or node identifier, that would make correlation even easier. This information should be visible directly in the run details view, and it should also be included in exported logs or audit events so we can correlate runs in external tooling like Splunk. Having it available via API would be a big plus too, since it would let us automate troubleshooting and reporting.

This would make it much easier to debug failures, spot patterns, and understand whether issues are tied to a specific worker or host, especially in environments that rely exclusively on private workers.

Status: ⬆️ Gathering votes3 comments

Log in to comment and vote

Comments3

  • Amaranth Dew

    •

    Aug 27

    It would be really, really helpful if the worker metadata tags (which for AWS show the worker’s EC2 instance ID) were visible after a worker is removed. Even if it just shows up in the worker pool view as a terminated worker. I need the EC2 instance ID to check logs/metrics and figure out why a worker might have crashed

  • Bronze Rice

    •

    Feb 25

    We also have a similar need as we were trying to track down a failed worker run that we thought might have been dispatched to Graviton (arm64) based compute and then failed.

    It would be great to report/record the worker pool on the run result view for easier debugging.

  • Black Breeze

    •

    Dec 15, 2025

    Thanks for raising this! The ID of the worker is available in the logs for the “preparing” stage (note the initial “Ground Control” logs). Would that work for you?