Live-site rollout

How live-site changes stay controlled.

Staged rollout, visible rollout state, change windows, scoped promotion, rollback and recovery, and audit-backed overrides—on a live warehouse site.

Staged rollout

Roll out in controlled stages

Rollout moves in steps: checks after each step, optional automatic rollback when health slips, and a fast return to a last known-good setup when you need it.

Start with a smaller site slice first.
Promote only when rollout checks pass.
Roll back fast when the outcome isn’t holding and you need a clean recovery.
Edgentik rollout flow: staged steps, checks, and rollback to a last known-good setup.

Staged rollout with visible checks and rollback to a last known-good setup.

Live-site panel: rollout status, slice taking changes first, health outcome, scope, and rollback controls.

Visible rollout status, rollout scope, and rollback controls in one live-site panel.

Visible rollout state

Make the rollout state inspectable

Teams trust changes more when rollout stage, health outcome, freeze state, and overrides sit together in one live view.

See which slice of the site is taking changes first right now.
See what blocked promotion and what passed health checks.
Review the same audit trail that captures rollout, rollback, and override decisions.
Change windows

Protect the pilot window without blocking controlled exceptions

A live rollout stays stable when change windows protect the environment and audited overrides stay visible when there is a real operational reason.

Window activeBlock risky changes while the site is under evaluation.
Reason requiredOverrides capture who acted, what changed, and why.
Audit-backedExceptions stay in the same trail as rollout activity.
Diagram: change windows during pilot and an audited path when an override is needed.

Change windows quiet the site during evaluation; audited overrides stay available when the job truly requires them.

Live-site rollout scope: flags, lanes, and staged promotion with checks before wider impact.

Flags and rollout-scope controls roll out new capability only when the site is ready.

Feature flags

Control rollout scope with flags and staged promotion

Flags are another lever for scoped promotion: turn on only what the site needs, where it needs it, after the right checks pass.

Scope capabilities by site, lane, or stage.
Promote with health checks and visible rollout state—widen impact deliberately.
Audit each enable, disable, and override—same trail as rollout and rollback.
Guided recovery

Recovery should stay inside the change workflow

When a rollout degrades, the recovery path stays visible: symptom-first triage, guarded actions, and a return to a known-good setup without guesswork.

Guided diagnostics keeps degraded and unknown states actionable with a clear path back to steady operation.

Guided recovery: symptom-first triage, health-informed rollback, and return to a known-good setup in the same rollout.

Rollback and recovery stay in the same rollout you widened—triage, then return to a known-good setup.