Product Fleet Operations — observability¶
Target
This page describes intended behaviour or structure. No implementation is claimed unless a section is explicitly marked Current implementation or Evidence.
| Metric | Service | Instrument | Dimensions |
|---|---|---|---|
connectsoft.factory.fleet.incidents_open |
FleetIncidentService | updown | tenant, productId, severity |
connectsoft.factory.fleet.time_to_acknowledge |
FleetIncidentService | histogram | edition, severity |
connectsoft.factory.fleet.time_to_resolve |
FleetIncidentService | histogram | edition, severity |
connectsoft.factory.fleet.offboardings |
FleetIncidentService | counter | tenant, state |
connectsoft.factory.fleet.products |
FleetInventoryService | updown | tenant, edition |
connectsoft.factory.fleet.instances |
FleetInventoryService | updown | productId, state, region |
connectsoft.factory.fleet.version_skew |
FleetInventoryService | gauge | productId |
connectsoft.factory.fleet.host.inbox_lag |
FleetOperationsHost | gauge | module |
connectsoft.factory.fleet.rollouts |
FleetRolloutService | counter | tenant, kind, state |
connectsoft.factory.fleet.rollout_duration |
FleetRolloutService | histogram | kind |
connectsoft.factory.fleet.rollout_halts |
FleetRolloutService | counter | tenant, reason |
The platform's signals¶
| Signal | Why it matters |
|---|---|
connectsoft.factory.fleet.products, instances by state, version_skew |
the shape of the fleet and how far instances lag the current release |
connectsoft.factory.fleet.rollouts, rollout_duration, rollout_halts |
how safely and how fast the fleet moves |
connectsoft.factory.fleet.incidents_open, time_to_acknowledge, time_to_resolve |
the managed-service commitment per edition, measured |
connectsoft.factory.fleet.offboardings |
the exit path is working and within its window |
Objectives and alerts¶
Product objectives live in SloService with instances as subjects; the edition's acknowledgement targets are objectives on this platform's own incident metrics; AlertRuleService fires on on-call gaps, halted rollouts and overdue off-boardings. Product-level spans carry connectsoft.factory.product_id, instance_id, rollout_id, incident_id.