Your developers don't ship code straight to production, and your updates shouldn't go out that way either. Approve a product once. Every new release then flows through your rings, and each ring opens only when the devices before it are healthy, right away or at a time you schedule. Windows and Linux, servers and desktops.
- Boot
- Logs
- Health
- 98%
- Boot
- Logs
- Health
- 98%
- Boot
- Logs
- Health
- 98%
- Owner
The CI/CD ideas you already trust, applied to updates
| CI/CD concept | In Automation Workflows |
|---|---|
| Pipeline | An Automation Workflow for an update type: Microsoft updates, third-party Windows apps or Linux packages |
| Trigger | The products you approve. Each new release enters the pipeline on its own. |
| Stages | Rings, each with its own device group and timing |
| Canary | The test ring, which always goes first |
| Gates | Boot verification, event log analysis, customizable health signals and success thresholds, plus a manual approval on any ring you choose |
| Deployment window | Scheduled promotion: once a ring passes its gate, the next one can wait for the time you set, so testing runs all week and production deploys on Saturday |
| Release | The approved updates, installed in the order you approved them |
| Dry run | Planning mode |
| Rollback | A ring that misses its threshold stops the pipeline and flags the devices that got the update. You roll back a run when you choose. |
| Pipeline chaining | Workflow Orchestration |
| Build board | The Kanban board: To Do, In Progress, Testing, Done |
Gates that look at the devices, not just the calendar
- Boot verification: did the device come back up after the update?
- Event log analysis: new errors, service failures or crashes after installation.
- Customizable health signals: service status, registry values, running processes, network reachability, defined per workflow.
- Manual approval per ring, for the rings where you want a person to sign off.
- Right away or on a schedule: once the ring before it passes, the next ring opens immediately or at the time you set, so testing can run all week and production can deploy on Saturday.
- Success thresholds per ring, with automatic pause on boot failures, critical service failures or application launch failures.
- Retry, resume or roll back: a run that stops flags the devices that already got the update, and you choose what happens next.
- Notifications by email, Teams or webhook when a ring is promoted or paused.
Know what your test ring doesn't test
Patchblox shows where production differs from your test ring, so a clean test run means what you think it means.
Before and after
Signed pre- and post-install scripts stop services, take backups or check prerequisites, with a continue-on-failure option for non-critical checks.
Automation Workflows and Intune Deployments
Intune's new Deployments feature, in public preview since September 2026, takes one specific version of one app, or one policy, and pushes it through timed rings. When the next version ships, an admin builds a new deployment and starts again. Automation Workflows do what Deployments do, staged rings, and add what they leave out: a continuous pipeline, gates that check the devices, and Windows updates and Linux on servers as well as desktops.
| Intune Deployments (preview) | Patchblox Automation Workflows | |
|---|---|---|
| What a rollout carries | One payload: a single app version or a single policy | A standing pipeline for the products you approve |
| When the next version ships | An admin wraps it in a new deployment and runs the rings again | It enters the pipeline automatically |
| What it rolls out | Win32 and Enterprise App Catalog apps, Settings catalog and Endpoint security policies | Microsoft updates, 600+ third-party Windows apps and Linux packages |
| Windows updates and drivers | Not among the supported payloads | Microsoft updates are a core update type |
| Platforms | Windows 10 and later | Windows and Linux, servers and desktops |
| How rings advance | On a schedule, by deferral time between rings | When gates pass: boot verification, event logs, your customizable health signals and success thresholds, plus a manual approval on any ring you choose. The next ring then opens right away or at a time you schedule. |
| When something breaks | Rings keep advancing on schedule unless an admin pauses or cancels by hand, so problems tend to surface as user reports | The rollout stops itself past the threshold you set and flags the devices that got the update. You retry, resume or roll back. |
| Before and after install | Not part of a deployment | Signed pre- and post-install scripts |
| Dry run | Not documented | Planning mode |
| Test coverage | Not documented | Test-ring coverage gaps |
Based on Microsoft's documentation for Deployments in Intune as of November 2026, while the feature is in public preview; Microsoft may change it before general availability. Microsoft Learn: Deployment plans and deployments
Where each fits: Intune Deployments are a good way to stage a single app or policy change. Automation Workflows run the continuous stream of Microsoft, third-party and Linux updates across servers and desktops. They work side by side.
Watch a release move through the rings
We'll show a workflow from approval to production, including a gate that stops a bad update.