n8n (workflow)
Community Edition
- pin data - execute the action once and cache response
- AI Agent - driven workflow
- filling table
see also
# Setup
# Client Setup ⮺
Auto-reconnect with tunnel using autossh + user service
$ systemctl --user status n8n-tunnel
$ journalctl --user -u n8n-tunnel -f
# Comparison
| Criteria | n8n | Camunda | Azkaban | Argo Workflows |
|---|---|---|---|---|
| Primary job | App/automation orchestration | Business process (BPM) | Batch job scheduling | Container/K8s workflow orchestration |
| Model | Visual node graph | BPMN 2.0 | Job DAGs | Steps/DAGs as Kubernetes CRDs (YAML) |
| Visual designer | Strong | Strong BPMN | Basic ops UI | UI to view/submit runs; design is mostly YAML/Git |
| Execution tracking | Run history | Process instances / tokens | Job/flow status | Per-workflow pods, logs, status in Argo UI |
| Triggers | Webhooks, cron, events | Messages, timers, APIs, humans | Schedules, manual | CronWorkflows, sensors (Argo Events), API, GitOps |
| Human tasks | DIY | First-class | No | No (not a BPM engine) |
| Where it runs | App server / Docker | App platform | Hadoop/JVM cluster era | Requires Kubernetes |
| Unit of work | Connector / JS node | Service task / user task | Shell/Hadoop/Spark job | Container (image + command) |
| On-prem | Yes | Yes | Yes | Yes (on your K8s) |
| Open source | Fair-code | Edition-dependent | Apache 2.0 | Apache 2.0 |
| Ops weight | Light | Heavy | Medium / aging | Medium–heavy if you already run K8s; heavy if you don’t |
| Best fit | Glue APIs & SaaS | Org processes & compliance | Legacy batch DAGs | CI/CD, data/ML pipelines, infra jobs on K8s |
Written on September 9, 2026, Last update on
workflow
automation
agentic-AI