Skip to main content
A ticket carries the exact release identity and the exact failing check, and most fixes come back as a release through your own channel. Meeting platforms change without notice, and when capture breaks the meetings are gone. A ticket opens with the machine state assembled locally — preflight report, admission-refusal message, release coordinates — and the complete payload shown before anything is sent, so triage starts from the exact release identity and the exact failing check. An air-gapped site delivers the same payload out of band.
Rung. There is no packaging command shipped for this yet. Today the payload is the artefacts you already have — the preflight report, the refusal message, and the station-report.yaml from step 5 — attached by hand.
Most fixes are releases. A ticket resolves as fixed in release X, and that release reaches you through the same channel, with the same evidence, promoted by your own gate. Severity shortens validation windows on our side; it never skips a stage. Advisories — a platform behaviour change, a fleet-wide defect, a withdrawn release — are published on the channel as signed content and surface on your next poll.

​
Optional outbound

Two classes, both off by default and both bounded by your telemetry rung, which is where the mechanism and the guarantees are written down.
  • Notifications — health degradation, failed sync, completed promotion, filed by Argo CD’s built-in notifications engine pointed at the support endpoint: configuration of a component you already run.
  • Usage reports — where your agreement defines usage-based terms. The metric is computable by you from your own deployment and the report is signed from your side, so reconciliation compares receipts.
Severity classifies what a finding means for capture in production, and maps to a delivery commitment in your subscription agreement. The mechanism is the same at every severity: evidence in, release out, your gate decides when it lands. Next: Operations · The station gate