Documentation

Appearance

Fix a remote site without a truck roll ​

A site device fails. The default instinct is a truck roll or a VPN negotiation with customer IT. Shared SSH keys and inbound ports burn days before anyone touches the fault.

Without Dataplicity ​

Support waits on network change requests. Engineering keeps a spreadsheet of bastion hosts. Every incident invents its own path from “something is wrong” to “we fixed it.”

With Dataplicity ​

  1. Start from the signal. Open the alert or incident with device context already attached - organisation, site or customer when mapped, severity, and first-observed time. Acknowledge so others do not duplicate work.
  2. Gather evidence. Search or tail logs scoped to that device (and class/customer when present). Compare a healthy peer if you have one. Keep the query and findings on the incident.
  3. Use interactive access only when needed. Remote shell or Wormhole reach the device through the outbound agent - no customer firewall change. Prefer the least invasive action; verify recovery with an independent check, not only a successful command.
  4. Close the loop. Resolve with cause, action, and remaining risk. Promote repeatable checks into monitors or scheduled tasks so the next failure is quieter.

Outcome ​

MTTR drops. Truck rolls are reserved for true hardware or site work.

Next ​

Related path: Internal IoT operations