​

Enterprise Observability & Audit

Note: Core observability (Analytics Engine, basic traces, Logpush examples) is available in the open core. The full production-grade compliance, SIEM integration, long-term immutable audit, and tenant-aware pipelines described here are part of the commercial Enterprise layer.

​

Pillars

1. Workers Traces (automatic)
   • End-to-end visibility across Service Bindings, DO RPC, Queues, Workflows, external fetches.
   • No code changes needed for basic tracing.
2. Logpush (Enterprise strength)
   • Push Workers Trace Events, HTTP requests, WAF events, Queue events, etc.
   • Destination: R2 (primary), or external SIEM (SentinelOne, Splunk, etc.).
   • One-click R2 setup + advanced options (custom fields, raw vs transformed headers, daily partitioning).
   • High volume support on Enterprise.
3. Tail Workers
   • Sample, filter, enrich, or redact logs in real time before they go to Logpush or storage.
   • Example: add tenant context, hash sensitive fields, drop health checks.
4. R2 as Immutable Audit Store
   • Combined with R2 Event Notifications → trigger Workflows for post-processing (indexing, compliance checks, retention policies).
   • Jurisdiction controls for data residency.
   • Lifecycle + Object Lock for immutability.
5. Analytics Engine + Dashboard
   • Keep high-cardinality real-time metrics (latency, fills, risk events).
   • Tenant-tagged datasets.
6. Workflows + Observability
   • Every Workflow step can emit structured events.
   • Workflow execution history is durable and auditable.

​

Example Logpush Job (R2)

{
  "name": "hoox-enterprise-audit",
  "dataset": "workers_trace_events",
  "destination_conf": "r2://hoox-audit-logs/{DATE}?account-id=...&access-key-id=...&secret-access-key=...",
  "output_options": {
    "field_names": ["EventTimestamp", "EventType", "RayID", "Worker", "Outcome", "TenantId", ...],
    "timestamp_format": "rfc3339"
  },
  "enabled": true
}

​

Tail Worker Example (Tenant Enrichment + Redaction)

// workers/tail-audit/src/index.ts
export default {
  async tail(events: TraceItem[], env: Env) {
    for (const event of events) {
      // enrich
      const tenant = event.scriptTags?.find(t => t.startsWith("tenant:"))?.slice(7) || "unknown";
      
      // redact secrets
      const logs = (event.logs || []).map(l => ({
        ...l,
        message: l.message?.replace(/secret:[a-z0-9]+/gi, "secret:[REDACTED]")
      }));

      await env.AUDIT_QUEUE.send({
        tenant,
        ray: event.rayID,
        logs,
        // ...
      });
    }
  }
};

​

Compliance Use Cases

• Trade reconstruction: Pull all events for a clientOrderId or traceId across Workers, DOs, and Queues from R2.
• Kill switch audit: Every activation + reason + affected positions logged immutably.
• Regulatory reports: Scheduled Workflow reads from R2 + D1, produces signed PDF via Browser Rendering, stores result + manifest in R2.
• Anomaly detection: Feed Logpush stream (or sampled Tail) into AI Gateway + Vectorize for real-time risk signals.

​

Retention & Cost

• R2 is dramatically cheaper than traditional logging at scale.
• Use lifecycle rules: hot (30d) → cold → delete or archive after N years.
• Enterprise customers often keep 7+ years for compliance.

​

Integration Points

• Existing trackAnalytics in shared can be extended to also emit to the audit path.
• All new Workflows and DOs should emit structured auditEvent payloads.
• Dashboard (Enterprise) should have “Audit Explorer” views querying R2 (via Worker or R2 SQL if available) + Analytics Engine.

​

Related

• architecture.mdx
• multi-tenancy.mdx — tenant tagging in logs
• Root wrangler.jsonc Enterprise profiles will include Logpush + Tail bindings.