First-day onboarding

A 6-step path to confidently operating DRMS Jamaica from your first login through to running a real response.

Your first 30 days

Operator navigation

The left sidebar is grouped into eight collapsible sections. They expand on click and remember your preference.

Top bar

• Crumb — shows your current location
• Search — coming soon: type any keyword to find an event, report, or person
• Account chip — your name + role; click for profile
• Sign out — clears your token; safe to use on shared workstations

Reading the dashboard

The dashboard shows you "what is happening right now" in eight tiles. Knowing each one cold means you can read the state of Jamaica in 10 seconds.

Tile-by-tile

Live Map walkthrough

The Live Map is your situational-awareness display during any active incident. Every layer can be toggled independently.

Layer groups

• Hazards — earthquakes (USGS live feed), hurricanes (NHC cones), tsunami evacuation zones, storm surge tiers Cat-1/3/5
• Operational — sensors, critical assets, active alerts, dispatch incidents, citizen reports
• Civic — emergency shelters, population centres
• Humanitarian — WASH failures, drone captures, missing-person last-seen, triage tags

Reading triage colours

5+ RED tags within 6 hours automatically raises a Mass Casualty Incident (MCI) critical alert.

Managing disaster events

A "disaster event" is the container that ties everything else together. Get the event lifecycle right and the rest of the platform behaves correctly.

Lifecycle states

How auto-attribution works

When a new damage report comes in:

1. System looks for any event in monitoring/response/recovery status
2. If one of them has the report's parish in its parish list, tag the report to that event
3. Otherwise tag to the most recent active event
4. If no active event exists, the report is filed unattributed

Damage assessments

Two assessment modes — rapid (post-event 24-72 h) and detailed (insurer/ODPEM submission).

Rapid assessment workflow

Detailed assessment

Detailed mode unlocks construction type, repair cost estimate, agricultural acreage damaged, livestock lost, infrastructure type (hospital/school/gov/bridge/utility), telecom status. Use this for insurer submission and Sendai reporting.

Dispatch & incidents

Incidents are operator-logged events that require a response — fires, traffic accidents, search-and-rescue calls. They live in the items table with category prefix em_ or incidents.

Creating an incident from a call

Shelter operations

Shelters have a roster (occupants checked in) and a capacity (maximum). The system auto-alerts at 80% and 95% occupancy.

Check-in workflow

Capacity alerts

Generating SitReps & ICS forms

Every active event can produce four printable documents on demand, all sourced from the same snapshot endpoint.

Available reports

Workflow

Family Tracing & Reunification (FTR)

Aligned with ICRC standards. Tracks missing persons, found persons, and search-for-family records, with automatic matching.

The three record types

• Missing — family has lost contact with someone; family member submits the record
• Found — operator records a person they encountered (in a shelter, at a hospital) whose family is unknown
• Searching — a person is alive but searching for their family (often used after evacuation)

Public submission

Citizens can register a missing person at drmsja.org/ftr-public.html — no login required. Each submission flows into the FTR queue for operator review.

Auto-spike alert

When more than 5 "missing" records are filed in 24 h, the system raises an FTR spike alert (6 h cooldown). Treat this as a mass-displacement indicator — open the FTR queue and start matching against shelter rosters.

WASH water quality testing

Aligned with Sphere standards. Field teams test water at shelters and community taps; failing tests automatically raise an alert.

Sphere thresholds

Health surveillance — WHO IDSR

Syndromic disease surveillance. Field clinics report case counts per parish per syndrome; anomalies trigger outbreak investigation.

Syndromes tracked

• Acute diarrhoea · Acute respiratory · Fever of unknown origin · Rash + fever · Vector-borne (dengue/zika/chik) · Post-disaster injury · Mental health presentation

Anomaly threshold

Cases > 2× the baseline rate (specified per report) flag the row as an anomaly and raise a health alert. Treat as suspected outbreak; coordinate with MoHW.

Mutual aid (EMAC)

Cross-jurisdiction resource requests — CARICOM partners, CDEMA, USAID/BHA, partner states.

Status lifecycle

submitted → acknowledged → matched → in_transit → delivered · or declined / cancelled at any step

Aging

Resource lifecycle

Once a resource request is matched to inventory or a donor, this view tracks every state transition.

States and SLAs

The kanban board flags cards in red when they're past SLA.

MIRA — Multi-cluster Initial Rapid Assessment

IASC standard. First-72-hour rapid assessment across 10 humanitarian clusters using INFORM severity (1=minimal, 5=catastrophic).

The 10 clusters

1. Food security & livelihoods
2. WASH
3. Health
4. Nutrition
5. Protection
6. Shelter & NFI
7. Education
8. Logistics
9. Early recovery
10. CCCM (Camp Coordination)

When to file

Within 72 h of incident onset. Multiple teams can each file their own row per cluster — the system shows the peak severity per cluster on the event page.

Drone / UAV imagery

Aerial captures from authorised UAV flights. Each capture is geo-tagged and damage-rated.

Upload flow

Auto-actions

Hurricane tracking — NHC

Live tracks pulled every 30 min from the US National Hurricane Center CurrentStorms.json feed.

What we capture per advisory

• Storm position (lat/lng) and intensity (kt)
• Classification (TS, H1–H5)
• Forecast cone (GeoJSON polygon)
• Forecast track (line)
• Pressure, movement vector

The Live Map renders the cone, track and current position automatically when the layer is on. Event pages show all active storms within 1500 km of Jamaica.

CCRIF parametric triggers

CCRIF SPC pays out automatically when objective thresholds are crossed. We watch them in real time so finance leadership knows when payout becomes eligible.

Thresholds watched

Each crossing records a ccrif_triggers row and raises a critical alert. Acknowledge each trigger after IC has confirmed it's a real threshold breach (not noise).

Triage tags (START method)

Simple Triage and Rapid Treatment — the field-standard mass casualty triage system. We track each tag with location, vital signs and disposition.

The four colours

• RED — Immediate: life-threatening but salvageable injury; transport priority 1
• YELLOW — Delayed: serious injury, stable, can wait briefly
• GREEN — Minor: walking wounded; ambulatory
• BLACK — Expectant/Deceased: not currently treatable or already deceased

MCI threshold

Donations ledger

Cash and in-kind contributions tied to events. Every $1M JMD milestone is auto-logged to the event timeline.

What to capture

• Donor name + organisation
• Cash amount + currency, or in-kind item + estimated JMD value
• Channel (bank / mobile_money / cheque / in_person / shipped)
• Restricted? (donor specified how it must be used)
• Acknowledgement sent? Receipt issued?

Reporting

Donations totals appear on the event page Standards section. Use the donation ledger CSV export for finance reconciliation.

Sendai Framework indicators

UN Office for Disaster Risk Reduction reporting — A through F indicators.

What we compute

Export at /api/sendai/indicators.csv with optional ?event_id + ?from + ?to filters.

Alert engine

Every alert in DRMS Jamaica goes through the same pipeline: capture → enrichment → severity → notification.

Severities

Sources

• Sensor thresholds (rain, river, wind, surge)
• CCRIF parametric triggers
• NHC hurricane proximity
• USGS earthquakes (M5.5+ within 3° of JM)
• Citizen reports (auto-classified by keywords)
• WASH water-quality failures
• Health surveillance anomalies
• FTR spike detection
• MIRA severity ≥ 4
• Drone aerial damage (major/destroyed)
• Mass Casualty (5+ RED triage tags)
• Shelter capacity (80% / 95%)
• Sensor silent failure (no readings 6h/24h)
• Cross-domain anomaly engine (2.5× baseline)

Anomaly detection

Cross-domain anomaly engine. Runs every 15 minutes, compares current 1-hour activity against a rolling 7-day baseline.

What it watches

• Call volume — anomaly_calls
• Citizen reports — anomaly_reports
• Damage assessments — anomaly_damage
• Alert engine itself — anomaly_alerts (a flood of alerts is itself a disaster signal)

Threshold

Current ≥ 2.5× baseline AND current ≥ 3 (small-N filter). Severity scales: 5× or more → critical, otherwise warning. Cooldown 6 h per signal.

Webhooks for partners

CDEMA, ODPEM, partner state EOCs and other systems can subscribe to receive real-time alerts.

How to register a partner

Payload format

{
  "event": "alert.created",
  "alert": {
    "id": 12345,
    "type": "ccrif",
    "severity": "critical",
    "title": "🌀 CCRIF TC wind threshold breached",
    "message": "...",
    "parish": "St Thomas",
    "lat": 17.91, "lng": -76.42,
    "event_id": 42,
    "created_at": 1748023200000,
    "source": "drmsja.org"
  }
}

Signature header

If a secret is set on the endpoint, each request includes X-DRMS-Signature: sha256=<hex> and X-DRMS-Timestamp. Partners should verify the signature against the raw body before trusting the payload.

Retries

Failed deliveries (5xx or network error) retry up to 3 times with exponential back-off (30 s, 5 min, 1 h). Permanent failures are logged in the delivery history per endpoint.

Auto-actions catalog

The system performs these actions automatically without operator intervention. Knowing them prevents "wait, why did this happen?" confusion.

Failure modes & recovery

Production-grade reliability means anticipating what breaks. Here's the catalog.

External feed failures

Internal failures

Backups & data retention

SQLite database at /home/myvirtualoffices.com/jims-backend/data/jims.db.

Backup strategy

• Daily snapshots — taken at 02:00 JST by host cron
• Retained: last 30 days locally, last 12 months in off-host storage
• Audit log system_audit is append-only; nothing is ever deleted

Manual backup before risky changes

sudo systemctl stop jims-backend
cp /home/myvirtualoffices.com/jims-backend/data/jims.db \
   /home/myvirtualoffices.com/jims-backend/data/jims.db.backup-$(date +%F)
sudo systemctl start jims-backend

Restore procedure

sudo systemctl stop jims-backend
mv /home/myvirtualoffices.com/jims-backend/data/jims.db \
   /home/myvirtualoffices.com/jims-backend/data/jims.db.corrupt
cp /path/to/backup.db /home/myvirtualoffices.com/jims-backend/data/jims.db
sudo systemctl start jims-backend

User & role management

Roles

Creating a new operator

Audit log

Every POST/PATCH/DELETE against a sensitive entity is recorded immutably in the system_audit table.

What's logged

• User id and name
• Action (created/updated/deleted)
• Entity type and id
• Before/after snapshot (where applicable)
• Summary message
• Timestamp + source IP

Querying

Open Standards & export → Audit log. Filter by user, entity type, or date range. Export to CSV for compliance review.

Subscriber alerts — "My Alerts"

Citizens subscribe via a free DRMS Jamaica account and manage their preferences in the subscriber portal.

How it works

• Citizen signs up at /portal-login.html with email + phone verification
• They draw or pick "monitoring zones" — the geofence polygons that define where they want alerts
• Per-zone preferences for alert types (hurricane, flood, earthquake, water-quality, etc.)
• Alerts delivered to the in-app "My Alerts" page and (where enabled) via email / Web Push
• Geofence-driven — alerts only fire when an event intersects a zone the subscriber owns

Subscriber count and active zones appear in the admin dashboard. Manage subscribers from Admin → Users.

Frequently asked questions

The dashboard shows nothing — is the system broken?

First check if you're signed in. If your token expired, you'll see empty tiles. Sign in again. If still empty, ping https://drmsja.org/api/health in the browser — should return {"ok":true,...}.

I can't sign in — "Invalid credentials"

Confirm caps-lock is off. If you're sure the password is right, your account may be suspended — contact your admin. Last-resort: ask admin to reset your password.

Why didn't my colleague get the alert I just raised?

Operator alerts go to the dashboard in real time. They do NOT go to email/SMS unless your colleague is also a subscriber. To page someone, use the SMS/email broadcast tool (admin only).

Reports I'm filing aren't showing up under any event

Either no event was active at the time, or the parish in your report doesn't match any active event's parish list. Open the event and add your parish to the parish list, or create a new active event.

The map is showing stale data

Hard-refresh the browser tab (Ctrl+Shift+R or Cmd+Shift+R). Caching on the JS bundles is aggressive by design — service workers serve stale data while fresh data loads.

A sensor I know is working shows "silent"

The sensor reading hasn't reached the database within the last 6 hours. Check the sensor's uplink (cellular signal, power, antenna). The most common cause is power outage at the sensor site.

Where are uploaded drone photos stored?

Server-side at uploads/drones/, served at https://drmsja.org/drone-images/<file>. They survive backend restart. They are subject to the standard backup schedule.

Can I delete a report I filed by mistake?

Reporters cannot delete reports — but can edit them. Moderators/admins can delete; that deletion is recorded in the audit log.

What's the difference between an "incident" and an "event"?

An incident is one specific occurrence — a single house fire, a single car accident. An event is the umbrella container (Hurricane X, Earthquake Y) that aggregates many incidents.

Glossary & acronyms

Standards alignment

DRMS Jamaica implements or aligns with the following international, regional, and national frameworks.

UN / international

• UN Sendai Framework for Disaster Risk Reduction 2015–2030 — indicators A through F computed and exportable as CSV
• UN INFORM Severity Index — 1–5 scale used in MIRA assessments
• IASC MIRA — Multi-cluster Initial Rapid Assessment, 10 clusters supported
• OCHA HXL — Humanitarian eXchange Language tagging in open-data CSV exports
• ICRC Family Links — FTR record structure aligned with ICRC standards
• WHO IDSR — Integrated Disease Surveillance and Response, syndromic case reporting
• Sphere Standards — WASH water-quality thresholds enforced

Regional

• CDEMA Comprehensive Disaster Management (CDM) — framework alignment
• CCRIF SPC — parametric trigger thresholds watched in real time
• CARICOM EMAC — mutual aid request lifecycle

National (Jamaica)

• ODPEM — coordination protocols
• MoHW — health surveillance integration
• NWA — road closures and infrastructure damage
• JPS — power outage data

Technical / interoperability

• NIMS / ICS — Incident Command System forms 201/202/209 generated
• OASIS CAP 1.2 — Common Alerting Protocol feed exposed
• OpenAPI — REST APIs documented
• HMAC-SHA256 — outbound webhook signing
• WebSocket (Socket.IO) — real-time event push
• Service Worker / PWA — offline-capable progressive web app

Security

• Argon2/bcrypt password hashing
• JWT signed tokens for sessions
• Role-Based Access Control (RBAC) — admin / moderator / reporter / subscriber
• Append-only system audit log
• Rate limiting on public endpoints
• CORS allow-list
• HTTPS enforced via host (LiteSpeed + Let's Encrypt)

Keyboard shortcuts

Troubleshooting

Backend service not responding

sudo systemctl status jims-backend
sudo journalctl -u jims-backend -n 50 --no-pager
sudo systemctl restart jims-backend

Database appears corrupted

# Stop service, check integrity
sudo systemctl stop jims-backend
sqlite3 /home/myvirtualoffices.com/jims-backend/data/jims.db "PRAGMA integrity_check;"
# If not "ok", restore from latest backup
sudo systemctl start jims-backend

NHC hurricane data isn't updating

Check journalctl -u jims-backend | grep nhc. Common cause: NHC API rate-limited the host IP. Wait 30 min — the next poll cycle will succeed.

WebSocket / real-time updates stopped

Refresh the page. Socket.IO reconnects automatically but may need a manual nudge after long sleep states. If persistent, check that the LiteSpeed proxy is allowing /socket.io to upgrade to WebSocket.

"Origin not allowed by CORS"

You're accessing the API from a domain not in CORS_ORIGINS. Add the new domain to .env and restart.

Sensor showing readings but no alert raised

Open the sensor record; confirm thresholds are set. The monitoring engine only fires alerts when reading crosses the threshold direction (e.g., wind > 60 kt). A reading of 60 kt exactly does not fire — set thresholds to your action threshold minus a small safety margin.