Skip to content
GeoBridge logo GeoBridge logo GeoBridge Docs

Privacy & Telemetry

Geobridge is designed around “no raw query persistence.” Telemetry is scoped to operational signals while omitting customer payloads and secrets.

What we collect

Section titled “What we collect”
  • Structured logs: event name, trace ID, component metadata, anonymized account/key IDs, error summaries. Request bodies and raw geocoding queries are never logged.
  • Metrics: bounded tags (route, outcome, cache status, job hash) and numeric counters for edge requests, bulk processing, and webhooks. Tag values are truncated to avoid free-form payloads.
  • Alert telemetry: aggregated counts/ratios for webhook failures, bulk errors, cache misses. No customer identifiers beyond hashed job IDs.

Controls

Section titled “Controls”
  • Data minimisation: telemetry helpers drop null/undefined fields and refuse to serialize request bodies or webhook payloads.
  • Access: Workers AE datasets are limited to the SRE group; alerts redact customer identifiers.
  • Redaction tests: automated tests assert no raw query strings enter logs; webhook retry auditing avoids payload storage.
  • Webhooks: signed with HMAC SHA-256 using caller-provided secrets; failures are retried with backoff and summarized in job status.

Customer responsibilities

Section titled “Customer responsibilities”
  • Use distinct webhook secrets per integration.
  • Avoid embedding sensitive data in job record IDs or texts; they may appear in NDJSON outputs you store.
  • Configure retention-appropriate TTLs when submitting bulk jobs if you need shorter windows.
Section titled “Related links”