Skip to content

Local Dashboard

The local dashboard is a built-in web UI served automatically by the SDK on port 4242 (configurable). It is identical across all SDKs — the same interface whether you run Node.js or Python.

Cloud mode

In cloud mode, the local dashboard is not started. Metrics are visualized in the APIForge cloud dashboard instead. See Cloud Mode.

http://localhost:4242

No installation, no separate server, no configuration needed.

Panels

Health Score

A single 0–100 score summarizing the global health of your API, computed from four dimensions:

DimensionWeight
Availability (2xx rate, last 24h)30%
Performance (P90 vs historical baseline)30%
Stability (absence of anomalies, last 7 days)25%
Quality (proportion of active, healthy endpoints)15%

A score above 85 is healthy. Below 60 indicates a problem that needs attention.

Latency chart

Time series of P50, P90 and P99 latency per endpoint. Click any route in the table to focus the chart on that route.

Available time ranges: 24h / 7d / 30d.

Routes table

All instrumented routes, sorted by request volume. Columns:

ColumnDescription
RouteParameterized pattern — e.g. GET /users/:id or GET /users/{user_id}
RequestsTotal calls in the selected time range
P50 / P90 / P99Latency percentiles in milliseconds
Avg sizeAverage response body size in bytes (Content-Length), when available
Error ratePercentage of 4xx + 5xx responses
StatusOK, SLOW, DEGRADED, or DEAD

Insights panel

Automatically generated alerts — no configuration required. Filter chips let you narrow the list to a specific insight type: ANOMALY, DRIFT, DEAD, PERF, or UNTRACKED.

See Automatic Insights.

Disabling the dashboard

js
app.use(apiforge({ dashboardPort: 0 }))
python
app.add_middleware(ApiForgeMiddleware, dashboard_port=0)

Custom port

js
app.use(apiforge({ mode: 'local', dashboardPort: 9090 }))
// Dashboard → http://localhost:9090
python
app.add_middleware(ApiForgeMiddleware, dashboard_port=9090)
# Dashboard → http://localhost:9090

Security note

The dashboard has no authentication in local mode. Do not expose port 4242 to the public internet. Use a firewall rule or SSH tunnel if you need to access it remotely:

bash
ssh -L 4242:localhost:4242 user@your-server
# Then open http://localhost:4242 locally

Released under the MIT License.