MDM rollout overview
pavri-endpointd ships as a signed OS-native package and is rolled out via
the customer's existing MDM. This page is the entry point; per-vendor runbooks
live alongside:
- Jamf Pro — macOS
- Microsoft Intune — Windows
- Kandji — macOS
- VMware Workspace ONE / Omnissa — macOS + Windows
- Linux config management — Ansible, Chef, Puppet
Every runbook follows the same seven-step structure:
- Prerequisites — Pavri tenant + admin role; vendor admin role; signed package.
- Token mint — admin walks to
/admin/endpoints → Add Endpoint → Mint enrollment token. Tokens are single-use per endpoint, 24h TTL. - Substitution markers — every importable artifact in
infra/mdm/<vendor>/uses${ENROLLMENT_TOKEN}and${ORG_ID}substitution markers. Replace before import. - Package signing — what the vendor expects (Developer ID for macOS / Authenticode
for Windows / signed
.debor.rpmfor Linux) and the current CAP-3 status (CAP-3.7 item 1 Windows-Intel and CAP-3.8 macOS ES are external-dependency gates). - Staged rollout — observe (1 week, 5% fleet) → soft_enforce (1 week, 25%) → enforce
(gradual to 100%). The rollout uses existing
protection_profile+ endpoint-attribute filtering; the MDM only delivers the binary. - Rollback — how to roll back via the vendor and the dashboard-side
observe-mode emergency flip. - Post-install verification — Run the verification block in §7 of your vendor's
runbook to confirm the daemon is loaded and the native-messaging manifests are healthy.
Then confirm the endpoint shows up at the
/admin/endpointspage.
Continue to your vendor's runbook for the concrete steps.
Linux note: Ubuntu .deb install/enroll/observe is live-verified. RPM packages
have structural checks plus a scheduled/manual Fedora and RHEL-family live
validation harness. Linux fleet rollout is configuration-management based rather
than classic MDM.