Approve deployments with evidence, gates and rollout-aware visibility
Deployment approval cards, risk gates and live rollout status where Argo and Kubernetes are configured. Failed provider actions stay visible in the timeline and outbox. AI does not deploy to production on its own — promotion happens only after human approval.
A gate between ready and promoted
The AI surfaces the status and the proposal; a person decides whether to promote.
- 01Gate opensA deployment reaches an approval gate
When a change is ready to promote, a deployment approval card opens — surfacing what would be promoted and the risk attached to it.
- 02Status in viewRollout status, where it is configured
Where Argo and Kubernetes are connected, FixControl shows rollout status so the approver decides with the live picture, not a guess.
- 03Risk gatesRisk gates hold the risky promotions
Higher-risk promotions pause for an explicit human decision. The gate is the point where a person — not the AI — chooses to proceed.
- 04Promote / statusPromote on approval; abort-aware status
Where Argo is configured for it, an approved decision promotes a paused rollout. Abort and rollback status stays visible where Argo reports it — reading status and acting on it stay distinct.
Promotion you can govern and audit
Each capability is described honestly — what's available depends on what's configured.
Deployment approval cards
Promotions surface as approval cards that state what would change and the risk involved, so the person deciding has the context in front of them.
Risk gates before promotion
Risky promotions pause at a gate for an explicit human decision, rather than flowing through automatically.
Argo / Kubernetes status where configured
Where those providers are connected, FixControl reads and shows rollout status — so the gate reflects the real state of the deployment.
Promote or read status — where configured
Where Argo is configured for it, an approved decision promotes a paused rollout; otherwise FixControl reads and surfaces status, including abort and rollback status where Argo reports it. What's available is shown honestly per capability.
Failed provider actions stay visible
A provider action that fails surfaces in the operational timeline and the outbox, instead of disappearing — so a stuck or failed promotion is visible, not silent.
Audit trail of the decision
Who approved a promotion, from which channel, on what evidence and with what consequence is recorded — the deployment decision is auditable after the fact.
No production promotion without a human
Common questions
Does the AI deploy to production by itself?+
What does FixControl actually do at deploy time today?+
What happens if a provider action fails?+
Is Argo or Kubernetes required?+
Is the deployment decision recorded?+
Can a risky promotion be forced through automatically?+
See deployment governance in action
Walk through approval cards, risk gates and rollout status on a live demo.