DevOps / GitOps — 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.delivery.host.inbox_lag |
DeliveryHost | gauge | module |
connectsoft.factory.delivery.promotions |
EnvironmentPromotionService | counter | tenant, environmentClass, state |
connectsoft.factory.delivery.promotion_lead_time |
EnvironmentPromotionService | histogram | environmentClass |
connectsoft.factory.delivery.rollbacks |
EnvironmentPromotionService | counter | tenant, environmentClass, automatic |
connectsoft.factory.delivery.proposals |
PipelineExecutionService | updown | tenant, state |
connectsoft.factory.delivery.pipeline_runs |
PipelineExecutionService | counter | tenant, state |
connectsoft.factory.delivery.pipeline_duration |
PipelineExecutionService | histogram | tenant |
connectsoft.factory.delivery.merge_lead_time |
PipelineExecutionService | histogram | tenant |
connectsoft.factory.delivery.releases |
ReleaseService | counter | tenant, state |
connectsoft.factory.delivery.publication_failures |
ReleaseService | counter | tenant, kind |
The platform's signals¶
| Signal | Why it matters |
|---|---|
connectsoft.factory.delivery.proposals by state, merge_lead_time |
how much Factory-authored code reaches a merge, and how fast: the agent-success signal for code changes (§51) |
connectsoft.factory.delivery.pipeline_runs, pipeline_duration |
the system of record's health as the Factory sees it |
connectsoft.factory.delivery.releases by state, publication_failures |
releases stuck as Candidate name the component that was never published |
connectsoft.factory.delivery.promotions by outcome, promotion_lead_time, rollbacks |
the four delivery measures (lead time, frequency, failure rate, time to restore) per environment class |
Where they go¶
The Observability platform's DashboardService consumes proposal, run, release and promotion events into run timelines and project overviews; SloService can define objectives on promotion failure rate and time to restore; AlertRuleService fires on PromotionFailed, RollbackCompleted and publication failures. Spans follow the Factory attributes with connectsoft.factory.proposal_id, pipeline_run_id, release_id, promotion_id added.