Observability
Vorian Connect surfaces real-time operational metrics — message throughput, auto-verification rates, processing latency, delivery status, and edge agent health — on a single dashboard. Interface engineers and lab managers get continuous visibility into how the system is performing without logging into servers or querying databases.
How It Works
- The gateway exposes a metrics endpoint on port
9091and publishes real-time events through a Redis-backed Server-Sent Events (SSE) hub - The dashboard polls the observability API at
/api/observabilityfor a snapshot of current metrics: total messages, outcomes by verification status, processing latency, payload sizes, and ACK code distribution - Message outcomes are broken down by verification status — auto-verified, pending, flagged, rejected, and parse errors — with counts and visual breakdowns
- Per-analyte throughput shows which tests are generating the most volume, helping labs understand their instrument mix at a glance
- Edge agent heartbeat status is monitored continuously — the topology page shows online, degraded, and offline agents with per-agent queue depth sparkline charts
- SSE events fan out to all connected dashboard clients within a tenant, so every user sees the same real-time data without polling overhead
What's Included
- Message throughput dashboard: Total messages processed, broken down by verification outcome. Auto-verify rate shown as a percentage — labs can track how much of their workflow is hands-free
- Processing latency: Average and per-message latency from receipt to verification completion, displayed in milliseconds. Helps identify slow instruments or backlog conditions
- Payload metrics: Average message payload size in kilobytes. Useful for capacity planning and identifying unusually large messages that may indicate configuration issues
- ACK code distribution: HL7 acknowledgment codes (AA, AE, AR) grouped by verification status to show which outcomes are being communicated back to instruments
- Per-analyte volume: Throughput broken down by analyte code — see which tests dominate the lab's workflow and identify seasonal or shift-based patterns
- Edge agent health: Heartbeat monitoring with online/degraded/offline status. Per-agent queue depth shows how many messages are buffered locally waiting for cloud delivery
- Real-time SSE events: Tenant-scoped Server-Sent Events pushed through Redis pub/sub. Dashboard clients receive live updates for messages, delivery status changes, and agent heartbeat changes
- Prometheus-compatible metrics: The gateway exposes a
/metricsendpoint on port 9091 in Prometheus text format for integration with existing monitoring infrastructure
Where It Fits
Observability is the operational heartbeat of Vorian Connect. It sits at the top of the dashboard navigation and is designed to be the first screen a lab manager checks in the morning. The metrics answer the core operational questions: Is traffic flowing? Are results auto-verifying? Are any agents offline? Is the system keeping up with message volume?
The observability page is complementary to the Trace Inspector. While the Inspector is for deep-diving individual messages, observability is for scanning system health at a glance. Both are tenant-scoped — each lab sees only its own metrics. The SSE hub ensures that all users within a tenant see consistent, up-to-date data without browser refresh.
Technical Details (for evaluators)
The SSE hub (package sse) maintains per-tenant channels with fan-out to all connected dashboard clients. Events are published through Redis for cross-instance distribution — when the gateway processes a message, it publishes an event to Redis, and every gateway instance's SSE hub forwards it to connected clients. The hub supports TLS-secured Redis connections via REDIS_TLS=true for production deployments.
The observability API at /api/observability returns a snapshot with total message counts, verification outcome breakdowns (grouped by ACK code and verify status), processing latency averages, payload size averages, auto-verify rate, and per-analyte throughput. The dashboard polls this endpoint and renders color-coded outcome bars — green for auto-verified, yellow for flagged, red for rejected, slate for pending, orange for parse errors.
The Prometheus metrics endpoint on port 9091 exposes standard counters and gauges for message throughput, latency histograms, delivery success rates, and edge agent connection counts. Labs already running Prometheus/Grafana can ingest Vorian metrics alongside their existing infrastructure monitoring without additional tooling.