Skip to content
StratumStratum Docs

Readiness, settling and alarms

Readiness

Section titled “Readiness”

Readiness answers “can this cloud do its job right now?”. Every finding is recomputed from current state on each read, says what to do, and clears by itself once the gap is closed.

Terminal window
stratumctl admin readiness                  # findings and their fixes; exits 1 when not ready
stratumctl admin readiness --wait 30m       # wait until ready, for scripts
stratumctl admin readiness -o json
Severity Means
blocking An operation cannot be performed. The cloud is not ready.
warning It works, but a foreseeable case will fail.
info Nothing is wrong: a deliberate state worth knowing.

A cluster has findings of its own (formation, storage paths, Ceph, joins): see Check readiness and Troubleshooting a cluster.

Common findings on a single node:

Finding Fix
NoStorageClass, NoDefaultStorageClass Create a pool and a class, mark one default.
NoImages Add a base image or import one.
NoAlertChannel Add a notification channel so alarms reach someone.
TimeNotSynced Check NTP reachability. TimeSourceLocalClock (info) means an offline install keeps its own clock on purpose.
AdminPasswordDefault The admin still has the install-time password: sign in and choose one.
SingleInstallAdmin Grant platform-admin to a second person.
SubstrateNearReservation The platform uses over 90% of its reserved memory: raise the host reservation.
UpdateAvailable, SecurityUpdateAvailable A newer release is on the mirror. See Upgrades.

Console: Admin › Substrate › Overview.

Readiness findings in the console.
Readiness findings in the console.

Settling

Section titled “Settling”

Right after an install, an upgrade or a reboot, the cloud reaches ready within a minute or two while many background loops are still on their first pass. Until all of them have converged once, or 15 minutes have passed, the cloud is settling:

  • the console shows Starting: N of M instead of warnings;
  • warnings and firing alerts are held back, so a fresh start does not page anyone; blocking findings are never held;
  • afterwards the phase is ready, or degraded with what never converged.

Alarms

Section titled “Alarms”

The Platform Health pack ships alarms for the platform itself: readiness blocked for 10 minutes (cluster-not-ready), backups, upgrades, security updates, storage pools and more. Firing alarms show under Admin › Substrate › Alerts.

Alarms reach people only through a notification channel. Wire the operator webhook; reading the URL from stdin keeps it out of shell history, and further lines are request headers (Authorization: Bearer ...):

Terminal window
printf '%s\n' "https://hooks.example.com/stratum" | \
  stratumctl admin observability operator-webhook set --url-stdin
stratumctl admin observability operator-webhook get

Self-check

Section titled “Self-check”

stratumctl admin selfcheck runs an end-to-end check of the platform’s own services and credentials, including backup freshness.