Today
One deployable backend, PostgreSQL persistence, structured logs, auth, role-aware endpoints, and clean project boundaries.
This page turns the documented architecture into a visual system map. It shows how the mobile app, API, data platform, realtime layer, financial records, audit trail, and founder visibility work together today, and how the platform is designed to grow without forcing a premature microservice split.
This is the high-level founder picture. Customers move from location-aware service discovery into curated technician matching, technician-card selection built around the kind of service experience they want, booking, live arrival coordination, and in-service visibility. The API stays as a modular monolith, while the platform keeps clear seams for payments, notifications, reporting, realtime delivery, and future scale upgrades.
The platform is being shaped so technician fit, customer experience choice, card positioning, operations, finance, live coordination, and compliance reassurance are explainable later without a rewrite.
This is the refined customer and technician flow. It starts with service intent and technician fit, then moves through card-based technician choice, booking, acceptance, live arrival, and in-service coordination. It shows where the API, realtime updates, reliability records, and founder visibility all meet around one booking.
Licensing and required compliance checks are handled as platform qualification gates first. Customers can still open booking detail or an on-demand credential view to confirm license and required proof when they want additional reassurance. The exact details shown should vary by jurisdiction, but the default trust view should cover active status, expiry context, and experience-building signals that help the customer feel confident.
The platform is not relying on one giant log table. Operational logs, audit history, business metrics, and realtime telemetry are separated so finance, trust, support, and founder questions can be answered cleanly.
Requests, failures, integrations, and background-job activity. Good for diagnosing runtime problems.
Admin approvals, overrides, trust decisions, and financial adjustments. Good for founder and support review.
Booking funnel, technician response time, completion rates, payout status, and support-safe reporting.
Presence and active-booking GPS sessions. Bounded retention, separate from normal logs, and tied to legitimate workflows.
Option 2 is not “many databases now.” It is “explicit responsibilities now, specialized stores when needed.” This diagram shows which store owns which kind of truth.
The platform is intentionally staged so the team can gain reliability and observability first, then add new infrastructure only when a measured workload makes it worthwhile.
One deployable backend, PostgreSQL persistence, structured logs, auth, role-aware endpoints, and clean project boundaries.
Add append-only booking events, audit events, financial ledger entries, correlation flow, and transactional outbox.
Add SignalR for booking propagation, presence, and active-booking location sharing backed by jobs and in-process events.
Add Redis for multi-instance SignalR coordination and hot presence state, not for financial or booking truth.
Add read-side/search or telemetry stores only when reporting, discovery, or GPS load starts competing with core transactions.