D2 Automation Knowledge
A systems view of separating control, webhook ingress and workflow execution in a queue-mode n8n architecture.
Direct answer
Queue mode separates accepting/orchestrating work from executing it. Redis provides queue coordination, workers consume execution jobs, PostgreSQL stores durable n8n state, and webhook processors can be separated from the editor/control plane. The benefit is not 'more containers'; it is independent scaling and failure isolation between traffic intake, control-plane activity and execution capacity.
Engineering model
Ingress → queue coordination → worker execution → durable state; scale each plane according to its bottleneck
01 / Design rule
The editor/API process should not compete unnecessarily with high-volume workflow execution. Keeping responsibilities separate makes capacity planning and incident diagnosis clearer.
02 / Design rule
Redis coordinates queued jobs and worker consumption. Durable business state such as conversation ownership, order state or audit evidence belongs in the appropriate database or system of record rather than being inferred from queue state.
03 / Design rule
A healthy worker design minimizes unique local state. New workers can be added or restarted while shared configuration, encryption key compatibility, database and queue coordination keep execution consistent.
04 / Design rule
Adding workers does not fix slow downstream APIs, database contention or workflows dominated by serial waits. Observe queue depth, execution duration, failure rate and external rate limits before deciding which plane needs capacity.
Implementation checklist
FAQ
No. Queue mode is an execution architecture. High availability also depends on control-plane topology, database/Redis availability, load balancing, secrets, backups, monitoring and tested recovery procedures.
No. A smaller single-instance workload can be simpler and easier to operate. Queue mode is justified when execution isolation, concurrency or scale requirements outweigh the added infrastructure complexity.
Evidence standard
D2 publishes these boundaries explicitly. The methodology page explains what evidence is required before a system is described as implemented, validated or production-backed.
Read D2 evidence methodologyRelated system evidence
Apply the framework