PulseWatch

Self-hosted reliability dashboard. Probes, SLOs, public status pages — without the SaaS bill.

🔗 Live demo: pulsewatch-ui.onrender.com  |  📊 Status page: pulsewatch-ui.onrender.com/#/p/demo

---

---

Why

Better Stack costs $29/month. Datadog costs $300/month. Uptime Kuma is great but has no SLOs, no REST API, no config-as-code. PulseWatch fills the gap: SRE-grade tracking with one docker compose up.

Features

• HTTP probes with configurable assertions — status code, latency threshold, body regex, JSON path
• SLO tracking — availability, error budget burn rate, projected exhaustion over 7/30-day windows
• Incident autodetection — opens when availability drops below target, closes on recovery
• Public status pages — 90-day historical bars, incident timeline, custom slug (/p/<slug>)
• YAML config-as-code — define probes, SLOs, and status pages in pulsewatch.yaml; import via API
• Real-time dashboard via SignalR — no polling, no F5
• REST API + OpenAPI — full API surface at /scalar/v1
• Multi-tenancy — Organizations → Projects → Probes
• Self-instrumented — OpenTelemetry traces + /metrics Prometheus endpoint

Quick start

docker compose up -d
# Backend: http://localhost:8080
# Frontend: http://localhost:5173

Or run locally without Docker:

# Start Postgres
docker run --name pulse-pg \
  -e POSTGRES_PASSWORD=dev -e POSTGRES_DB=pulsewatch \
  -p 5499:5432 -d postgres:16

# appsettings.Local.json (gitignored):
# { "ConnectionStrings": { "Postgres": "Host=localhost;Port=5499;Database=pulsewatch;Username=postgres;Password=dev" } }

dotnet run --project src/PulseWatch.Api
cd client && npm install && npm run dev

Backend: http://localhost:5035 · Frontend: http://localhost:5173 · Scalar: http://localhost:5035/scalar

Configuration as code

version: 1
project:
  name: My Services
  slug: my-services
probes:
  - name: API
    url: https://api.example.com/health
    interval: 30s
    assertions:
      - status: 200
      - latency_p95_ms: 500
slos:
  - probe: API
    target_availability: 99.9
    window: 30d
status_pages:
  - slug: demo
    title: My Services Status
    probes: [API]

curl -X POST localhost:5035/api/v1/yaml-import \
  -H 'Content-Type: text/yaml' \
  --data-binary @pulsewatch.yaml

Create a probe manually

ORG=$(curl -s -X POST localhost:5035/api/v1/organizations \
  -H 'Content-Type: application/json' \
  -d '{"name":"My Org","slug":"my-org"}' | jq -r .id)

PROJ=$(curl -s -X POST localhost:5035/api/v1/organizations/$ORG/projects \
  -H 'Content-Type: application/json' \
  -d '{"name":"My Project","slug":"my-project"}' | jq -r .id)

curl -s -X POST localhost:5035/api/v1/projects/$PROJ/probes \
  -H 'Content-Type: application/json' \
  -d '{
    "name": "API Health",
    "url": "https://api.example.com/health",
    "intervalSeconds": 30,
    "assertions": [
      { "type": "StatusCode", "operator": "Equals",   "expectedValue": "200" },
      { "type": "LatencyMs",  "operator": "LessThan", "expectedValue": "500" },
      { "type": "JsonPath",   "operator": "Equals",   "expectedValue": "ok",
        "jsonPathExpression": "$.status" }
    ]
  }'

Architecture

ProbeScheduler (5 s tick)
    │
    │  Channel<ProbeJob> — bounded 1000, DropWrite
    ▼
ProbeWorker ×4 (concurrent)
    │  HTTP probe + assertion evaluation
    │
    ├─ HealthCheck row ──┐
    └─ OutboxMessage row─┤  single transaction
                         │
                    OutboxRelay (200 ms poll, FOR UPDATE SKIP LOCKED)
                         │
                    SignalR hub → browser (live dashboard)
                    IMemoryCache invalidation → status page snapshots

RollupRefresher (60 s)
    └─ REFRESH MATERIALIZED VIEW CONCURRENTLY (health_check_1m/1h/1d)

SloCalculator (60 s)
    └─ reads health_check_1h/1d → writes SloMeasurement + auto-opens/closes Incidents

Assertion engine: ProbeWorker dispatches to StatusCodeEvaluator, LatencyEvaluator, BodyRegexEvaluator, or JsonPathEvaluator. All assertions must pass for IsSuccess = true.

Architecture Decision Records:

• ADR 001 — Transactional Outbox
• ADR 002 — Channel Pipeline
• ADR 003 — PostgreSQL Rollups vs TimescaleDB

Stack

Backend — .NET 10, ASP.NET Core Minimal API, EF Core 9, SignalR, PostgreSQL, Serilog
Frontend — React 18, TypeScript, Vite, TanStack Query, shadcn/ui, Tailwind, recharts
Observability — OpenTelemetry (traces + metrics), Prometheus /metrics, Scalar OpenAPI UI
Tests — xunit, FluentAssertions, Testcontainers (PostgreSQL), WireMock.Net, Playwright

Run tests

dotnet test

Unit tests: < 1 s. Integration tests (Testcontainers): ~30 s.

# Frontend smoke tests (Vite dev server starts automatically)
cd client && npm run test:smoke

# Full E2E including backend flows (requires running backend)
cd client && BACKEND_AVAILABLE=true VITE_API_URL=http://localhost:5035 npm test

Roadmap

• OIDC authentication (Keycloak)
• Slack/Discord webhook on SLO breach
• Distributed probing from multiple regions
• Synthetic transactions (multi-step, Playwright-style)
• Data retention policies
• On-call rotation

License

MIT