Skip to main content

Fleet Operations

Your fleet dashboard exposes three operations for managing your workers: Restart Worker, Force Restart, and Relaunch.

Restart Worker

Restarts the Browserless service on a single worker (same VM, same hardware). Use this when one worker is behaving unexpectedly while the rest of the fleet is healthy.

  • Terminates active sessions on that worker only
  • Other workers continue serving traffic
  • Does not affect IP addresses
  • Completes in seconds

Access this from the Production Workers table in your dashboard by selecting an individual worker.

Force Restart

Restarts the Browserless service on all workers simultaneously (same VMs, same hardware). Use this for fleet-wide service issues when you want a quick reset without provisioning new infrastructure.

  • May interrupt active sessions across all workers
  • Does not affect IP addresses
  • Faster than Relaunch
warning

If CPU or memory are consistently high and Force Restart doesn't resolve the issue, use Relaunch instead. Restarting the service on an overloaded machine will not clear the underlying resource pressure.

Relaunch

Provisions entirely new VMs and containers using a blue-green deployment. Your current workers continue serving traffic while new ones come online. Once the new fleet is healthy, traffic switches over and the old VMs are decommissioned.

Use Relaunch when:

  • Session times have been gradually increasing over days or weeks
  • Force Restart did not resolve the issue
  • A node has gone offline and traffic is not routing correctly

Image version upgrades are also executed as a Relaunch. Browserless coordinates the rollout with you.

What changes during a Relaunch

Changes
API tokenNo, stays the same
Worker IP addressesYes, new VMs get new IPs
Active sessionsNot interrupted (traffic continues on old workers during transition)
Dashboard metricsReset on the new workers
estimatedMonthlyUnitsUnreliable for several hours (insufficient history on new workers)

Duration: Approximately 5 to 10 minutes.

Update firewall allowlists after a Relaunch

Each worker has a fixed public IP shown beneath its title in the Workers section of your cluster page. When you Relaunch, these IPs change. If you or any third-party service uses IP-based allowlisting, update those lists after the new workers are online.

The new IPs appear in the Workers section once the Relaunch completes.