Skip to content

Fleet Rollout And Templating

What This Covers

  • per-site overrides
  • shared base config
  • staged rollout
  • rollback planning across more than one runtime

Start Here

Create and inspect managed local runtime projects before planning a wider rollout:

trust-runtime fleet runtime add --fleet-root ./fleet --name cell-1 --template simulate --json
trust-runtime fleet list --fleet-root ./fleet --json

The add command creates ./fleet/cell-1 and registers it in ./fleet/fleet.toml. Runtime names must be portable and unique. The command allocates distinct available loopback control/web ports and a secure random control credential; malformed hand-edited manifests are rejected before a runtime is added or controlled.

Managed runtime lifecycle commands are:

trust-runtime fleet runtime start --fleet-root ./fleet --name cell-1 --json
trust-runtime fleet runtime status --fleet-root ./fleet --name cell-1 --json
trust-runtime fleet runtime logs --fleet-root ./fleet --name cell-1 --lines 100 --json
trust-runtime fleet runtime stop --fleet-root ./fleet --name cell-1 --json

Only a failed connection is reported as stopped. Invalid local credentials or a connected endpoint that returns a rejected or malformed control response are errors, preventing an uncertain state from being mistaken for a safe restart.

What Success Looks Like

  • each site has one shared base profile plus only the overrides it needs
  • rollout order, rollback owner, and rollback trigger are written before rollout starts
  • at least one runtime is upgraded and verified before the fleet-wide step