Tech retweeted
🛠️ Every DevOps Engineer and SRE should know Vim.
When you're debugging a production issue over SSH, editing a config on a remote server, or troubleshooting a Kubernetes pod, Vim can save you valuable time.
You won't always have VS Code available. Sometimes, all you have is a terminal. ⚡
Bookmark this Vim cheat sheet to master the essentials:
✅ Navigate files like a pro
✅ Edit configs without leaving the terminal
✅ Search logs and replace text quickly
✅ Work efficiently on remote servers
Pro tip: Master dd, /pattern, :%s/old/new/g, and :wq first.
Small skills. Faster debugging. Better incident response.
♻️ Repost this to help another DevOps engineer or SRE!
#DevOps #SRE #Linux #Vim #Kubernetes
Tech retweeted
Kubernetes Pod Networking — explained visually.
Every Pod gets its own IP.
Pods can communicate across nodes.
Services provide stable access to changing Pods.
CNI plugins connect the cluster network.
Understand these 4 concepts, and Kubernetes networking starts making sense.
Save this for your next K8s revision. 🔖
#Kubernetes #DevOps #CloudNative
Tech retweeted
Payment succeeded. Shipping failed.
The customer has paid, stock is reserved, but delivery cannot be arranged.
How do you recover when each service has its own database?
This is where the SAGA pattern helps.
WHAT IS A SAGA?
A SAGA breaks one business operation into a sequence of local transactions.
Each service completes its own work and commits changes to its own database. If the workflow cannot finish, compensating actions address the effects of earlier successful steps.
Let’s use an online checkout.
THE SUCCESSFUL FLOW
1. Order service creates a PENDING order.
2. Inventory service reserves the product.
3. Payment service processes the payment.
4. Shipping service books delivery.
5. Order service marks the order CONFIRMED.
These are separate commits across services.
WHAT IF SHIPPING REJECTS THE BOOKING?
The payment and stock reservation have already succeeded.
For this workflow, recovery could be:
1. Refund the payment.
2. Release the reserved stock.
3. Mark the order CANCELLED.
These recovery operations are called compensating actions.
COMPENSATION ≠ DATABASE ROLLBACK
A refund is a new transaction. It does not erase the original payment.
Similarly, releasing stock is a new inventory operation.
Compensation follows business rules to reach a valid outcome. It does not necessarily restore every detail of the original state.
HOW ARE THE STEPS COORDINATED?
There are two common approaches:
• Choreography: Services react to events. Inventory publishes StockReserved, Payment responds, then Shipping responds to PaymentCompleted. There is no central coordinator. Simple flows can work well, but larger workflows become harder to trace.
• Orchestration: A coordinator requests each step, tracks outcomes, and directs recovery. The workflow is easier to follow, but the coordinator must reliably persist its progress.
WHAT MAKES A SAGA RELIABLE?
• Idempotency: Repeated commands must not cause duplicate charges or refunds.
• Durable state: A restart must not lose completed steps or pending recovery.
• Reliable messaging: A committed change must not silently lose the event that advances the workflow.
• Safe retries: A timeout does not prove an operation failed. Check the outcome before repeating a payment.
• Recovery handling: Compensation can fail too. Record progress, retry safely, and alert when recovery stalls.
SAGA allows intermediate states—such as payment completed while shipping is pending. It aims for eventual consistency and does not provide built-in global isolation.
Would you choose choreography or orchestration for this checkout—and why?
#SystemDesign #Microservices
Tech retweeted
Top 10 System Design Interview Questions
1. Design a URL Shortener (Bitly)
2. Design WhatsApp / Chat App
3. Design Instagram / News Feed
4. Design a Distributed Rate Limiter
5. Design Uber / Ola
6. Design YouTube / Video Streaming
7. Design an E-commerce Order System
8. Design a Notification Service
9. Design Dropbox / Google Drive
10. Design a Distributed Job Scheduler
