AXIS SENTINEL CONTINUITYCore system
Protection is not enough. The organization must survive.
SENTINEL PROTECTS AND VERIFIES. PRESERVATION AND RESTORE STAY A SEPARATE MANDATE.
Continuity is how Sentinel carries an organization through an incident. Sentinel detects, contains, preserves, requests continuity action and verifies the recovered state. A separate recovery mandate maintains restore points and executes recovery.

- 01ObserveContinuous telemetry from every service and edge you connect.
- 02VerifyTelemetry checked at the gateway before it is trusted.
- 03DetectAnomaly, policy violation, and absent signal alike.
- 04ProtectPolicy enforced at the identity and tenant boundary.
- 05PreserveEvidence written down before it can be lost.
- 06RecoverContinuity coordinated, state restored, result verified. (Continuity operates here)
Recover returns to Observe
Continuity closes the loop: state restored and the result verified.
Separate systems
Two systems. Two mandates.
Continuity is how Sentinel carries an organization through an incident. It rests on a boundary: what Sentinel does, what Recovery does, and the narrow contract between them. The two are never merged.

Sentinel above and Recovery below, two separate systems. Sentinel: a pair of platinum rings and a column of cool light that watches and verifies. Recovery: a stack of broad plates with warm light leaking between them, which preserves and recovers. The contract: one narrow line between them carries a preservation request down and a restore point or a failure back. Sentinel does not itself restore; this shows the doctrine.
Sentinel protects and verifies. Its light comes from above: it watches, contains and checks what comes back.
- SentinelSentinel protects and verifies. Its light comes from above: it watches, contains and checks what comes back.
- Recoverypreservation and restore stay a separate mandate. It is a separate system below: layer on layer of restore points, kept quiet until they are needed.
- The contractBetween them, one narrow line: a preservation request goes down, and a restore point or a failure comes back. Sentinel does not itself restore; this is the doctrine.
AXIS Sentinel
Protects and verifies.
- Detects the incident and declares it
- Contains it at the boundary, through Guard
- Preserves evidence of what happened
- Requests preservation at the moment of detection
- Records Recovery's reported outcome in the Ledger
- Verifies the recovered state before an incident may close
Sentinel does not
- Perform backup, snapshot or restore
- Run recovery
The boundary is a contract
Sentinel does not itself restore. The outcome is recorded as an operator’s attestation of what Recovery reported, and is labeled as such.
a separate recovery system
Preserves and recovers.
- Maintains restore points
- Preserves critical state
- Executes recovery
- Returns a verified recovery result
A separate recovery mandate. It is not part of Sentinel, and Sentinel does not implement it.
The continuity view
What is protected, what returns first, what is checked.
The same facts the product keeps, in one view: the services that never stop, the order the organization comes back in, and the states a preservation request and a recovery pass through before an incident may close.
Protected, throughout
- Identity
- Evidence capture and Ledger writes
- Incident response
- Preservation requests
Safe Mode never holds these. Identity, evidence capture, incident response and preservation requests keep running while everything around them transitions.
Comes back, in order
- a separate recovery systemVerified restore pointHolds the restore point, reports the outcome
- a separate recovery systemRecoveryExecutes recovery from the restore point
- SentinelIntegrity verificationVerifies the recovered state
- SentinelReturn to operationCloses the incident, lifts Safe Mode
Checked before it counts
- Preservation requestedSent at detection, before remediation can overwrite state.
- Restore point confirmedRecovery's confirmation recorded, with who attested it.
- Preservation failedRecorded as failed. The incident stays open.
- Recovery not verifiedThe incident may not close yet.
- Recovery verifiedIntegrity checked. Return to operation.
An incident cannot close while a recovery-verification task is open. The gate records that verification was attested; it does not grade it.
The incident
The core survives while everything around it transitions.
Detection, preservation request, verified restore point, containment, recovery, integrity verification, return to operation. Each step belongs to one of the two systems. Preservation outcomes are recorded in the Ledger that Evidence keeps.
Illustrative demonstration
One incident, seven steps
Step 7 of 7
Return to operation
Sentinel returns the organization to operation. The incident may close and Safe Mode may be lifted, with authority and a recorded reason. The surrounding systems are operating again, the whole sequence is in the record, and the loop returns to Observe.
Sentinel detects. Watchtower or Guard raises the condition and an operator declares an incident; a playbook assigns the checklist. The protected core is untouched, and everything Sentinel needs to respond stays available.
Sentinel requests. At the moment of detection, before any remediation that could overwrite state, it opens a preservation request linked to the incident. The request is a record that Sentinel asked; the preservation itself is Recovery's. Safe Mode never holds this request.
Recovery preserves. It holds the restore point and reports the outcome. Sentinel appends that outcome to the Ledger before it marks the job confirmed, and says who attested it: today an operator, reporting what Recovery said. If the outcome is a failure, that is recorded too, and the incident stays open.
Sentinel contains. Guard halts spread at the boundary, and an operator may engage Safe Mode at the narrowest scope that covers the incident. High-risk operations are held; identity, evidence capture and incident response keep running. The surrounding systems transition while the core holds.
Recovery recovers. It executes recovery from the verified restore point, along its own recovery paths. Sentinel does not perform backup, snapshot or restore, and does not run recovery.
Sentinel verifies. The recovered state is checked against the incident's recovery criteria and the recovery-verification task is attested by an operator. An incident with that task still open cannot be closed. The gate records that verification was attested; it does not grade it.
Sentinel returns the organization to operation. The incident may close and Safe Mode may be lifted, with authority and a recorded reason. The surrounding systems are operating again, the whole sequence is in the record, and the loop returns to Observe.
Simulated states, written to show how the two systems divide the work. Nothing here reads or changes a live system. Sentinel does not itself restore.
Signal path
What enters it, and what leaves.
Signals in
- Declared incidents and their playbook tasks
- Safe Mode state
- Recovery's reported outcome, attested by an operator
- Recovery verification tasks
Continuity
RECOVER
Outputs
- Preservation requests linked to the incident
- Confirmed or failed outcomes in the Ledger
- A close refused until recovery is verified
- Continuity state for Command
Ten questions
Continuity, answered.
The questions every Sentinel system must answer, in the same order on every system page.
Continuity is Sentinel's recovery coordination system. Sentinel detects, contains, preserves, requests continuity action and verifies the recovered state. Preservation and restore stay a separate mandate.
- Sentinel protects and verifies. Preservation and restore stay a separate mandate. The two are never merged.
Incidents that need state preserved, each preservation request and its outcome, and the recovery-verification step that stands between an incident and its close.
Incident declarations and playbook tasks, Safe Mode state, and Recovery's reported outcome for each preservation request.
When to request preservation, at the moment of detection and before any remediation that could overwrite state, and whether recovery has been verified before an incident may close.
Open a preservation request linked to the incident's playbook task, record the confirmed or failed outcome to the Ledger, and refuse to close an incident while a recovery-verification task is still open.
Sentinel never performs backup, snapshot or restore. That is Recovery's mandate.
Design directionSentinel does not itself restore. A preservation outcome is recorded as an operator's attestation of what Recovery reported, and is labelled as such.
Preservation jobs with their status and who attested the outcome, the incident's timeline, and which recovery-verification tasks block its close.
Continuity follows Guard's containment, works from the controlled state Safe Mode holds, records every outcome to Evidence, and is meant to be exercised by Range recovery scenarios, which are not in the catalog yet.
If preservation cannot be confirmed, the job is recorded as failed and the incident stays open. The record of what was attempted outlasts the incident.
The close transition reads the checklist back on every attempt, so verification is enforced rather than remembered. Range does not yet have recovery scenarios.
Relations
Where it connects.
- Guard Receives fromContainment that precedes recovery
- Safe Mode Exchanges withRecovery proceeds from a controlled state
- Evidence Sends toPreservation outcomes recorded to the Ledger
- Range Receives fromRecovery scenarios (not yet in the catalog)
- Command Sends toPreservation jobs and what blocks an incident's close
- Separate recovery mandate Exchanges with, outside SentinelPreservation requests out; restore point and recovery result back
- Sends to
- Receives from
- Exchanges with
- Outside Sentinel
Failure and health
When it fails, it says so.
In operation
Restore point confirmed
- Preservation requested at detection
- Outcome recorded to the Ledger
When it fails or loses visibility
Preservation failed
- Job recorded as failed, with who attested it
- The incident stays open
- The attempt remains part of the record
Continuity states
- Preservation requestedSent at detection, before remediation can overwrite state.
- Restore point confirmed (lit: in operation)Recovery's confirmation recorded, with who attested it.
- Preservation failed (lit: on failure)Recorded as failed. The incident stays open.
- Recovery not verifiedThe incident may not close yet.
- Recovery verifiedIntegrity checked. Return to operation.
How its own health is verified
- Close refused while recovery verification is open
- Range: recovery scenarios are not yet in the catalog
- The gate reads the checklist on every attempt
The Sentinel system family
Ten systems, one institution.
Put Sentinel on guard
Choose the protection level built for where your company is today.
