Release 1.1.0

Summary

Highlights include quality-of-life improvements in the visual editor, faster iteration in the dashboard, and better stability on macOS hosts.

Released: 4 Feb 2025

Added enable/disable toggles for actions. The toggle sets condition: "false" behind the scenes, so advanced users can still hand-edit conditions in raw mode.
Map-style inputs (for example key/value field lists) now support in-place edits instead of delete-and-recreate workflows.
Transient job cancellation now targets the explicit job name instead of assuming index 0, preventing accidental clashes when multiple people test at once.
The run output view adds clearer guidance on the Capture output option and lets you turn off syntax highlighting when large payloads were bogging down browsers.
The raw job editor warns about unsaved changes by highlighting the Close raw editor button in yellow when edits have not been applied.

The dashboard surfaces a dedicated New job shortcut so builders can jump straight into authoring.
Metrics and logs screens add a back control tied to browser navigation, making it easier to return to the originating job or worker context.

Lifted the capture pipeline out of the print output. You can now capture events inside Run & Trace while still exercising your real output connectors.
Added capture-and-pass-through support for transient jobs so you can test an output end-to-end before deploying.
CLI jobs import gained an --update flag to rebuild job history in place, and simple import/export gained parity for round-tripping definitions.
Windows packaging now includes an MSI installer that registers Lyft Data as a service (via WinSW) and encourages running under a domain or service account instead of the local context.

Reduced file descriptor usage on macOS hosts where the default limits are strict.

Prevented filtered events from losing their attachments in traces so you can follow the full lineage in Run & Trace.
Resolved file input/output jobs that continued looping despite stop_reading_after: true; they now behave as one-shot jobs.
Fixed file store default behaviour: path is mandatory, and the visual editor no longer requires a bucket field when unnecessary.
Addressed Splunk HEC form defaults so only HEC token remains mandatory, with updated help text per docs review.
Several DSL help items (Files, Log Files, File Monitor) were renamed for clarity and filter how options were documented/adjusted, including the new script-based predicate.

Air-gapped licensing: if the server cannot reach the licensing service for an extended period the built-in worker may eventually stop accepting jobs. Upgrade to 1.1.1 or keep a Community Edition key ready as a fallback.
Large Run & Trace payloads can still feel heavy. Turning off syntax highlighting is the recommended workaround until the editor receives further optimisation.
Output failures triggered through /api/run may still report a generic “shutting down” message. Collect the worker logs when escalating to support.

No database migrations are required; roll workers and the server in place.
Restart the web UI after upgrading the backend so the new editor assets are loaded.
If you scripted transient job cancellation via the API, update calls to pass the job identifier returned by /api/jobs/run/list.