Step 6 - Diagnostic Service

Goal

Create a Service resource that exposes diagnostic-service HTTP through the default Router mesh. The alert dashboard on ops-b uses this bridge in step 8.

Scenario context

When an operator clicks Diagnose on an alert, the dashboard calls a runbook API on site-a. That microservice listens on container port 8080 (external 8081 on the Edgelet node). It does not use NATS.

Router bridges HTTP across Edgelet nodes without exposing factory ports on the public internet. This step creates the bridge; step 7 reads the assigned bridgePort.

YAML walkthrough

File: deploy/steps/05-service-diagnostic.yaml

diagnostic-service Service

---
apiVersion: datasance.com/v3
kind: Service
metadata:
name: diagnostic-service
tags:
  - pot-edge-patterns
spec:
type: microservice
resource: production-monitoring/diagnostic-service
targetPort: 8081
defaultBridge: default-router

Field	Value	Meaning
resource	production-monitoring/diagnostic-service	Application/microservice pair from step 3
targetPort	8081	External port on site-a (maps to container 8080)
defaultBridge	default-router	Platform Router instance
tags	pot-edge-patterns	Tag consuming agents for service discovery

Ensure the ops-b Edgelet node carries tag service.iofog.org/tag: pot-edge-patterns so Router routes traffic to this Service.

Why this design

NATS carries alerts and telemetry. HTTP runbooks use Router because diagnostic-service is a plain REST API. The pattern scales to any remote ops center: NATS for events, Router for request/response.

Deploy

potctl deploy -f deploy/steps/05-service-diagnostic.yaml

Verify

potctl get services
potctl describe service diagnostic-service

The describe output includes bridgePort. Save it for step 7 and step 8 (DIAGNOSTIC_URL).

Optional curl from a shell on ops-b after step 8 sets the URL:

curl -s "http://router.default.svc.bridge.local:<diagnostic-bridge-port>/diagnose/M002?faultType=high_vibration"

Common mistakes

• Deploying before step 3. The microservice production-monitoring/diagnostic-service must exist.
• Wrong targetPort. Use 8081 (external), not container port 8080.
• Missing Router tag on ops-b. Without pot-edge-patterns tag, cross-node HTTP routing fails with timeout or 502.
• Skipping step 7. You need bridgePort before deploying the dashboard Application.

Next step

Step 7: Note bridge port: read bridgePort from potctl describe service diagnostic-service.

See anything wrong with the document? Help us improve it!

Edit this page on GitHub!