Varqeni Help

The fixes Varqeni can run

Updated 11 November 2026

Varqeni can only run actions that are written into its code, reviewed and tested on injected faults.

AvailabilityAvailable in early access when your policy allows it. The autopilot covers the two pipeline fixes; the four infrastructure fixes need a named approval.

How it works

  • Resume a paused connector (for Connector paused). Varqeni reads the connector again first and resumes it only if it is still paused.
  • Restart a failed task (for Task still failed after an outage), only once the outage behind it has cleared.
  • Restart a host's node exporter (for Host unreachable, Infrastructure), after a named approval, only on hosts your administrator lists, only if the exporter does not answer.
  • Free disk space (for Disk filling): vacuum the journal, delete old rotated logs and clear the package cache. Never data or a live log.
  • Rolling restart (for a hung workload) and roll back a Deployment (for a bad rollout), only in the namespaces your administrator lists.
  • Start a stopped virtual machine (Infrastructure). Built, and switched off: it needs a named approval and its own access token.

What you should see: Every fix is checked afterwards, and Varqeni never retries on its own. The four infrastructure fixes passed 32 of 32 closed-loop fault trials in the Fortis lab, and the wrong fix was never sent.

What it can and cannot do

Can

  • Run six reviewed fixes after a named approval: resume a paused connector, restart a failed task, restart a metrics exporter, clear old logs, restart or roll back a deployment, and check each result.

Cannot

  • Retry a fix on its own, or start a stopped virtual machine today. That action is built and switched off.

Still stuck? Contact support or email info@fortisdatagroup.com.