GeoSentinel exposes three lightweight diagnostic endpoints: a basic liveness check that confirms the API process is running, a database connectivity probe that executes a round-tripDocumentation Index
Fetch the complete documentation index at: https://mintlify.com/danizd/GeoSentinel/llms.txt
Use this file to discover all available pages before exploring further.
SELECT 1 against PostgreSQL, and a development seed route that populates the database with a minimal set of known-good incidents. These endpoints require no request body and carry no side effects in the case of the health probes — making them safe to call from load balancers, uptime monitors, and CI pipelines.
GET /v1/health
Returns the service’s liveness status. This endpoint does not touch the database and will respond as long as the FastAPI process is alive. Use it as a shallow heartbeat check in orchestration environments (Kubernetes livenessProbe, load balancer health targets, etc.).
Response — 200 OK
Always
"healthy" when the API process is running and able to handle requests.The service identifier. Always
"geosentinel".GET /v1/health/db
Verifies database connectivity by executing SELECT 1 against the configured PostgreSQL instance via SQLAlchemy. Use this as a deeper readiness check when you need to confirm that the full data path is operational — for example, as a Kubernetes readinessProbe or in a deployment smoke test.
Response — 200 OK (connected)
Response — 200 OK (disconnected)
When the database query raises an exception, the endpoint still returns HTTP 200 but the body reflects the failure state and includes the raw error message.
"healthy" if the SELECT 1 query succeeded; "unhealthy" if an exception was raised."connected" on success; "disconnected" on failure.Present only when the database probe fails. Contains the string representation of the exception raised by SQLAlchemy.
GET /v1/seed
Inserts a fixed set of sample records into the database for development and testing purposes. The endpoint is idempotent with respect to source metadata — it checks for an existing SourcesMetadata row before inserting and skips the source insertion if records are already present. Incident records are always inserted on each call.
The following data is seeded on each invocation:
Source metadata (inserted only if no sources exist yet):
gdelt— GDELT Cloud Events v2 (media_derived)usgs— USGS Earthquake Hazards (sensor)firms— FIRMS NASA VIIRS/MODIS (sensor)acled— ACLED CC BY-NC 4.0 (field_reported)
| Category | Event Type | Location |
|---|---|---|
conflict | conflict_battle | 40.0°N, 3.0°W |
disaster_natural | earthquake | 35.0°N, 120.0°W |
wildfire | wildfire_hotspot | 45.0°N, 110.0°W |
status = "open" and include severity_max, severity_latest, confidence, source_count, and observation_count fields populated with representative values.