Notice when a product line silently splits in the field

Shipping with Alignment

Class alignment probes are rolling out with the Alignment feature on device classes. Until a dedicated how-to lands, configure and view results on the device class Alignment tab. This story describes the workflow intent within current product behaviour.

A firmware or ops lead owns a device class in the field. After months of incremental image changes, nobody knows whether “the class” is still one reality. Spot checks on a few units hide a silent minority until a rollout fails.

Without Dataplicity

Tribal knowledge. Surprises when half the fleet answers differently. No shared history of how alike the class has been.

With Dataplicity

1. Configure one class probe - a diagnostic shell command (default uname -srm for kernel/arch, or a command that matters for your product line).
2. Devices report samples. The platform fingerprints matching stdout and exit codes into response classes.
3. Watch the mix over time - majority share and distribution history on the Alignment tab.
4. Drill into the minority bucket - which devices, what excerpt, when probed.
5. Decide the next action offline - image fix, cohort tag, support playbook. Alignment is observational: it does not enforce desired state or auto-remediate.

Outcome

You see a split before customers force the issue.

Next

• Device classes
• Fleet jobs (when you act on what you learned)

Related: Field advantages, Internal IoT operations