Milestone 4: Fabric rollback completed

This is the fourth milestone checkpoint. The instructor will review the rollback workflow from this section before the final part of the lab.

✅ Compared the fabric graph before and after the project expansion — two different network states, side by side
✅ Used Time Voyager to stage a rollback to the pre-project approved state
✅ Reviewed the staged diff showing what would be removed, changed, and restored across the entire fabric
✅ Understood that rollback is a single system-wide operation, not a per-device manual undo
✅ Confirmed the live network was unchanged before deciding to commit or revert

The key idea in this section is not the rollback button itself — it is what the rollback represents.

In a conventional network, rolling back means identifying every device that was touched, finding the previous configuration state, and manually re-applying it — device by device, hoping nothing was missed.

In Apstra, the network is a system. Every committed state is a snapshot of that system’s complete intent. Rolling back does not mean undoing device configurations one by one. It means returning the whole system to a previous coherent state in a single operation — and previewing exactly what that means before you commit to it.

That changes how you think about risk. Changes become less permanent. Decisions become more reversible.

This is the last section of the lab. You will use Data Centre Assurance — Juniper’s cloud-based AIOps platform — to see what happens when you combine Apstra’s intent model with machine learning and Flow data across multiple sites.

You will need to sign up for the demo environment. The sign-up link and steps are at the start of the next section.

Ready? Let’s move on!