Skip to main content

PulsePoint Incidents

Not every dot on the map is a camera. PulsePoint is an external feed of emergency incidents (fires, medical, hazards) that operators want to see alongside the camera network. Bringing that feed in is a small, self-contained integration that lives in CameraOperation.

The one idea to hold on to​

PulsePoint incidents are pushed in over HTTP and written to two places at once — Cosmos for the platform's own record, and ESRI for the map — with a daily job to keep both honest.

PulsePointIncidentPush is a function-authenticated HTTP endpoint (POST/PUT) that upserts each incident. It's marked [Singleton] so overlapping pushes don't race each other into inconsistent state — incidents are processed one batch at a time.

Gotcha: two stores means two failure surfaces

Because an incident is written to both Cosmos and ESRI, the two can diverge if one write succeeds and the other doesn't. The daily cleanup is what eventually reconciles them — so a transient ESRI hiccup shows up as a briefly stale map, corrected on the next reconcile, rather than a hard failure. Don't assume the map and Cosmos are instantaneously identical.

Wrapping up​

PulsePoint is a contained external integration: an HTTP push upserts emergency incidents into both Cosmos and ESRI, serialized by a singleton to avoid races, with a daily production cleanup reconciling the two stores. It's independent of the camera pipeline — just another operational job that happens to share this app.

Next, where captured images and their metadata actually live, and how they're retired: Image Storage & Cleanup.