#Production Profiles

A production profile is the combination of tuning posture, host readiness, and operational verification that qualifies a deployment as production-ready. This page describes what production readiness means across the ByteOr stack.

#What Production Readiness Requires

Production readiness is verified, not assumed. The runtime reports what it was able to apply, and operators confirm the result matches expectations.

#Host Readiness

• doctor passes for the requested tuning profile
• Memory-lock limits are sufficient
• CPU isolation is configured (for isolated-core)
• RT scheduling permissions are available (for isolated-core)
• SHM placement is correct

#Runtime Readiness

• The spec validates and compiles cleanly
• The compile report shows no unexpected diagnostics
• The policy audit confirms approval coverage
• The capability contract matches the spec requirements

#Operational Readiness

• Heartbeats are flowing from the agent
• Snapshots are being captured
• Artifact uploads are succeeding
• The audit trail records deployment creation and status transitions

#IndexBus Production Requirements

IndexBus production profiles add transport-layer requirements:

• SHM files are on a tmpfs or hugetlbfs mount
• No stale SHM files from previous runs
• Router and merge infrastructure is correctly sized
• Lane capacities match the workload

#Enterprise Production Requirements

Enterprise adds governance requirements:

• Approval coverage for the target environment
• Policy audit records show clean verification
• Incident bundle export is functional
• Replay dry-run path is tested

#Cloud Production Requirements

Cloud adds control-plane requirements:

• Agent enrollment is complete
• Runtime API keys are distributed
• Config bundle is associated with the deployment
• Deployment status transitions are being recorded
• Artifact upload routes are reachable from agents

#Verification Workflow

1. Run doctor on the target host
2. Validate the spec with validate
3. Check compatibility with compatibility
4. Deploy through the approval-gated path
5. Confirm agent heartbeats and status
6. Verify effective tuning in deployment details
7. Test incident bundle export and dry-run replay