Deployments

GitLab builds it, LynxTrac lands it

GitLab pipelines end with a package; your fleet starts with an installer that has to land on machines that are sometimes off, sometimes behind hotel Wi-Fi, and occasionally mid-reboot. LynxTrac bridges that gap with artifact ingestion from GitLab Packages and a delivery engine built for unreliable endpoints.

Deployments run as state machines per device, which means a laptop that was closed during the rollout picks up where the fleet left off instead of being forgotten.

Package registry ingestion

Deployments pull from GitLab Packages directly, with scoped credentials for private projects.

Staged, gated delivery

Roll to a canary group first, hold for approval, then continue. Release templates make the second deployment as fast as the tenth.

Rollback as a plan, not a scramble

Each release carries its rollback path. Validation failure on a device triggers it automatically for that device, with the full trail in the audit log.

Setting it up

Add GitLab as an artifact source, supply a scoped token for the project or group, and pick the package the deployment should track.

Credentials are stored encrypted like every other platform secret, and one release workflow can mix GitLab with other artifact sources.

Artifact-level today: GitLab CI finishes by publishing a package, and LynxTrac picks up from there. Native GitLab CI triggers are planned but not shipped.

Related: Continuous deployments

See it connected to your own account

The free tier covers 2 servers forever, which is enough to wire this up and judge it on your infrastructure rather than ours.