DevOps revision -Part1
• localhost ≠ deployment
• Domains are human friendly, IPs run the internet
• Private IPs = intranet, public IPs = global reach
• VMs are cloud computers with real responsibility
• SSH keys > passwords
DevOps starts with systems thinking, not tools
Backend engineering isn’t about writing more code.
It’s about putting code in the right place.
Routes handle paths.
Controllers handle HTTP.
Services make decisions.
Workers execute.
Clear boundaries turn scaling into configuration — not refactors.
3/
Why this matters 👇
In automation platforms, triggers and actions are not harmless features.
They can call external APIs, modify data, trigger workflows, and access sensitive credentials.
2/
Worked across:
⚡ Triggers — how workflows are initiated
⚙️ Actions — execution logic & data flow
🔐 Credentials — secure storage & access
🧭 Workflow ownership & execution boundaries
1/
Backend architecture work — automation platform internals ⚙️
Spent time strengthening the foundation of an n8n-like system.
This layer controls how workflows, triggers, actions, credentials interact —and decides whether the platform scales safely or breaks under real usage
While building a workflow management backend inside a TurboRepo monorepo, I am focusing on API boundaries, user-scoped data access, service-layer architecture, and type safety with TS + Prisma
Big takeaway: backend engineering is about trust boundaries & clean system design.
Currently building an n8n-like workflow automation system.
Not a CRUD app — the real work is in:
• execution flow
• trigger → executor design
• modular node architecture
genuinely enjoying the systems-level challenges. 🚀
@kirat_tw#buildinpublic#backend#systemdesign