Security and Compliance¶

AsyncMQ provides runtime primitives; secure operation is your responsibility.

Core Practices¶

• run backends on private networks
• enforce TLS where supported
• use least-privilege credentials for workers/producers
• avoid putting secrets or sensitive raw PII in job payloads
• enforce idempotency and authorization in producer endpoints
• keep the default max_job_payload_bytes guard enabled unless producers are independently authenticated, rate-limited, and bounded upstream

Dashboard Security¶

• enable authentication (AsyncMQAdmin(enable_login=True, backend=...))
• use strong session/JWT secrets
• restrict dashboard exposure by network controls and identity provider layers
• keep dashboard CORS restricted to the same origin unless an exact operator origin is required
• never combine wildcard CORS origins with credentialed dashboard sessions
• leave mutation checks for the same origin enabled unless an upstream gateway enforces equivalent CSRF protections
• authenticated dashboard mutations require an explicit Origin from the same origin header by default

Logging and Data Handling¶

• do not log secrets from job args/kwargs
• control retention of completed/failed jobs (purge / cleanup jobs)
• ensure storage encryption requirements are met by your backend deployment
• prefer references to large objects instead of embedding bulk data in job payloads; public Queue and task enqueue APIs reject JSON-encoded payloads above max_job_payload_bytes

Compliance Notes¶

For GDPR/HIPAA/PCI-like requirements:

• keep payloads minimal (prefer references over full records)
• define clear retention and deletion workflows
• audit who can enqueue, inspect, retry, and delete jobs

AsyncMQ can be part of compliant architectures, but compliance is system-level, not library-only.