Add datasource adapters for telemetry and status #8

New issue

Open

opened 2026-06-18 16:12:49 +02:00 by vince · 0 comments

vince commented

2026-06-18 16:12:49 +02:00

Owner

Parent: #1

Goal

Add the first live-data layer for metrics and service health while keeping datasource definitions in the dashboard model and rendering logic generic.

Scope

Define datasource reference types in the model for static values, Prometheus queries, HTTP checks, site monitor badges/status endpoints, and weather/current-condition sources.
Implement server-side adapter interfaces for fetching datasource values.
Add timeout, cache, stale, unavailable, and error handling states.
Normalize metric values, units, thresholds, sparklines, and status severities into renderer-friendly data.
Return sparkline/time-series data in a shape that can feed uPlot where richer charting is needed.
Support configured fallback values for first-version placeholders.
Use msw in tests to mock Prometheus/status/weather APIs without coupling tests to live infrastructure.
Add tests/mocks for successful, stale, failed, and partially unavailable datasource responses.

Acceptance criteria

Telemetry cards can render live, cached, stale, and unavailable states from adapter results.
Service health/status values are resolved through datasource references, not component hardcoding.
Adapter failures do not break the whole dashboard.
Refresh behavior is configurable through the model or app settings.
No secrets or private tokens are exposed to client-side code.
Mocked datasource tests cover the main success and failure paths.

Out of scope

Full observability backend.
Long-term time-series storage.
Agent UI for configuring datasources.

Parent: #1 ## Goal Add the first live-data layer for metrics and service health while keeping datasource definitions in the dashboard model and rendering logic generic. ## Scope - Define datasource reference types in the model for static values, Prometheus queries, HTTP checks, site monitor badges/status endpoints, and weather/current-condition sources. - Implement server-side adapter interfaces for fetching datasource values. - Add timeout, cache, stale, unavailable, and error handling states. - Normalize metric values, units, thresholds, sparklines, and status severities into renderer-friendly data. - Return sparkline/time-series data in a shape that can feed `uPlot` where richer charting is needed. - Support configured fallback values for first-version placeholders. - Use `msw` in tests to mock Prometheus/status/weather APIs without coupling tests to live infrastructure. - Add tests/mocks for successful, stale, failed, and partially unavailable datasource responses. ## Acceptance criteria - Telemetry cards can render live, cached, stale, and unavailable states from adapter results. - Service health/status values are resolved through datasource references, not component hardcoding. - Adapter failures do not break the whole dashboard. - Refresh behavior is configurable through the model or app settings. - No secrets or private tokens are exposed to client-side code. - Mocked datasource tests cover the main success and failure paths. ## Out of scope - Full observability backend. - Long-term time-series storage. - Agent UI for configuring datasources.