Requirements
- Enterprise API access
- A FullHunt organization ID
- An HTTPS receiver that accepts JSON and deduplicates
idempotency_key - Python 3.10 or later with
requests
Delivery contract
Run the collector
deliver_enterprise_alerts.py
Receiver requirements
The receiver must makeidempotency_key unique. A network failure can occur after the receiver creates a ticket but before the collector receives the response. Retrying the same event must update or return the existing ticket, not create another one.
Checkpoint an alert only after a successful receiver response. For the first run, set FULLHUNT_BASELINE_ONLY=1 to record existing alerts without opening historical tickets.
Operating policy
- Use bounded date windows in scheduled jobs and retain overlap so late records are still collected.
- Keep ticket disposition, owner, severity, and closure state in the downstream system.
is_seenis platform data, not a ticket workflow. - Restrict alert messages and ticket data to the tenant identified by
organization_id. - Retry
429and temporary server failures with bounded backoff in a production worker. - Send malformed records to a dead-letter queue instead of silently marking them delivered.
