Skip to content

Microservices

Target Architecture — Final-State Design

This page describes the final-state microservice topology of the Observability & Feedback Platform. All services follow the ConnectSoft.Factory.Observability.{Service} naming convention and the Clean Architecture layering of the ConnectSoft microservice template.

The platform comprises eleven microservices, each owning one or more aggregate roots within a single bounded context. Services communicate through the canonical event envelope on Azure Service Bus (via MassTransit) and expose synchronous APIs only where a request/response contract is required. High-volume ingestion services (TraceService, MetricAggregationService) scale independently of the transactional services (IncidentService, FeedbackService).

Service Catalog

Microservice Responsibility APIs Events Aggregate Roots Store
ConnectSoft.Factory.Observability.TraceService Ingest and serve correlated distributed traces by traceId. GET /traces/{traceId} publishes TraceRecorded TraceRecord Application Insights
ConnectSoft.Factory.Observability.LogQueryService Govern tenant-scoped search over structured Serilog logs. POST /logs/search — LogRecordReference Log Analytics
ConnectSoft.Factory.Observability.MetricAggregationService Roll raw counters/histograms into queryable metric series. POST /metrics/query publishes MetricAggregated MetricSeries Application Insights
ConnectSoft.Factory.Observability.DashboardService Define and serve reusable multi-tenant dashboards. POST /dashboards — DashboardDefinition Azure SQL / PostgreSQL
ConnectSoft.Factory.Observability.AlertRuleService Manage alert rules and raise triggers. POST /alerts/rules publishes AlertTriggered AlertRule Azure SQL / PostgreSQL
ConnectSoft.Factory.Observability.IncidentService Manage the incident lifecycle with trace lineage. POST /incidents publishes IncidentOpened, IncidentResolved Incident Azure SQL / PostgreSQL
ConnectSoft.Factory.Observability.FeedbackService Capture durable feedback items from signals, humans, agents. POST /feedback-items publishes FeedbackItemCreated FeedbackItem Azure SQL / PostgreSQL + Blob
ConnectSoft.Factory.Observability.QualityScoreService Compute per-project / per-artifact quality scores. GET /quality/projects/{projectId} publishes QualityScoreComputed QualityScore Azure SQL / PostgreSQL
ConnectSoft.Factory.Observability.CostTelemetryService Attribute cost per tenant/project; detect anomalies. GET /cost/projects/{projectId} publishes CostAnomalyDetected CostSignal Azure SQL / PostgreSQL + Blob
ConnectSoft.Factory.Observability.SloService Define SLOs, track error budgets, detect breaches. — (internal) publishes SloBreached SloDefinition Azure SQL / PostgreSQL
ConnectSoft.Factory.Observability.TelemetryCorrelationService Stitch traces, logs, metrics, feedback into one view by traceId. — (internal) consumes all; emits correlation snapshots TelemetryCorrelation Application Insights + Azure SQL

Service Interaction Diagram

flowchart TB
    subgraph Ingest["Ingestion"]
        Trace["TraceService"]
        LogQ["LogQueryService"]
        MetricAgg["MetricAggregationService"]
    end

    subgraph Detect["Detection"]
        Slo["SloService"]
        AlertRule["AlertRuleService"]
        Dashboard["DashboardService"]
    end

    subgraph React["Reaction & Learning"]
        Incident["IncidentService"]
        Feedback["FeedbackService"]
        Quality["QualityScoreService"]
        Cost["CostTelemetryService"]
    end

    Correlation["TelemetryCorrelationService"]

    Trace -->|"TraceRecorded"| Correlation
    LogQ -->|"log refs"| Correlation
    MetricAgg -->|"MetricAggregated"| Correlation
    MetricAgg -->|"series"| Slo
    MetricAgg -->|"series"| AlertRule
    MetricAgg -->|"series"| Dashboard
    Slo -->|"SloBreached"| Incident
    AlertRule -->|"AlertTriggered"| Incident
    Incident -->|"IncidentResolved"| Feedback
    Trace -->|"runtime signals"| Feedback
    MetricAgg -->|"usage"| Cost
    Cost -->|"CostAnomalyDetected"| Feedback
    Feedback -->|"FeedbackItemCreated"| KP["Knowledge Platform"]
    Quality -->|"QualityScoreComputed"| KP
    Correlation -->|"correlated views"| Incident
    Correlation -->|"correlated views"| Quality
Hold "Alt" / "Option" to enable pan & zoom

Service Notes

  • TraceService is read-optimised over Application Insights. It does not store trace bodies in its own database; it queries App Insights and projects a correlated TraceRecord view keyed by traceId, enriched with the required telemetry dimensions.
  • LogQueryService is a governed query facade over Log Analytics. It enforces tenantId scoping and translates factory-domain filters (project, module, agent) into KQL.
  • MetricAggregationService runs both an API (ad-hoc queries) and is driven by the MetricAggregationWorker for scheduled rollups.
  • TelemetryCorrelationService and SloService have no public API surface; they are internal services driven by events and workers, exposed only through dashboards and the correlation views consumed by IncidentService and QualityScoreService.
  • FeedbackService and QualityScoreService are the learning services — their events are the platform's primary contribution to the Knowledge Platform improvement loop.

All services emit their own telemetry into the same OTEL/Serilog pipeline (see Observability) — the platform observes itself with the same tooling it provides to the factory.