Health and status
Two commands tell you how the instance is doing. Neither changes anything.
kindo statuskindo doctorStatus reports the installed version and whether the instance is healthy, then lists what isn’t fine: steps that didn’t finish, components on the wrong version or unhealthy, and a missing superadmin. An unhealthy component or a failed step makes status report that the instance isn’t healthy; a missing superadmin or a component on another version is a warning. A step left running when installing stopped shows as interrupted, and status names the install command that continues it. On a cluster with no instance, status names the next step: creating the config if none is stored, then installing. If the superadmin is missing, status prints the commands that create it, with the superadmin email from your config. To show only what’s recorded, without checking the cluster:
kindo status --no-liveThe health check looks for problems and, for each one, says how to fix it. It fails when it finds a critical problem, can’t reach the cluster, or finds no Kindo on the cluster; warnings alone don’t fail it. When it can’t read the Helm releases, the workloads, the Secrets they use, the jobs and sandbox pools, the records of earlier runs, or the superadmin, it warns that it didn’t check that area, and doesn’t report the area as fine. Fix the cause, then run it again. Run it after every upgrade and config change. It checks:
| Area | What it checks |
|---|---|
| Stored config | The config exists, can be applied, and holds every generated password and key, including the values Kindo can’t generate again |
| Versions | The Kindo CLI and the instance are on the same version |
| Helm releases | No Kindo component is stuck partway through installing, upgrading, or rolling back |
| Secret store | The store the config names is the one running, and it’s ready. An entry in the store that overrides a config setting is noted, not flagged |
| App secrets | The app secrets exist and are in sync, and none are left behind for an application that isn’t installed |
| Workloads | No workload refers to a Secret that doesn’t exist |
| Readiness | Kindo’s workloads are ready, no one-time setup keeps failing, and the standby sandbox pods for agents start and keep running. These are warnings |
| Public URLs | The URLs the applications serve match the ones the config produces, and won’t move at the next apply |
| Nango | Nango’s keys match the config, and the applications agree on them, including the webhook key |
| Hatchet | Hatchet accepts the applications’ token |
| Google sign-in | Its config settings are either complete or absent |
| Databases | Each service’s database user and database exist, and each service can sign in with the password in the config. A server the check can’t reach, one whose TLS is set to verify, or an instance not yet moved from 2026.09 shows as not checked |
| Superadmin | The instance has a superadmin |
| Last run | The last time you installed, upgraded or applied the config, every step finished, and no run skipped a check |
While another Kindo command is changing the cluster, the health check says which one.
