Sumit

Devops Engineer | platform Engineer | Agentic AI

[Sumit @sumiiitkumar]

Open to Work | DevOps / Platform Engineer

Hi everyone,

I’m Sumit Kumar, a Cloud & DevOps Engineer with 3 years of experience. I’m actively looking for DevOps Engineer or Platform Engineer opportunities. My expertise includes Kubernetes, AWS, Azure, Docker, Terraform, CI/CD, Linux, Python, Observability and AI Agents .

If your organization is hiring or you know of any relevant openings, I’d appreciate a referral or connection.

[email protected]

#OpenToWork #DevOps #cloud #PlatformEngineer #CloudEngineer #Kubernetes #AWS #Azure #Terraform #Docker #CICD #Hiring #AIAgents

[Aravind @aravind]

I again request the GoI. As these protests are resolved, India must:

- Ban social media for under 13
- Moderate social media for under 18
- Regulate social media companies
- Mandate a reporting tool under every post to GoI
- Setup a SM/M monitoring & response cell
- Study & research social media trends & Algo
- Take up propaganda on state media with countries
- Take up anti-Indian SM psyops with partners
- Monitor and control "finfluencers" farms
- Start units that can counter the infowar on India

And I hope India is using the current opportunity to study social media operations, organization, and algorithmic warfare of the adversaries on India.

To take the right corrective measures soon, understanding that adversaries and anti-india cabal have infiltrated SM companies with insiders.

So simple ad-hoc solutions like MPs, or even cyber armies, posting our narrative on Insta now is going to solve little, as algorithm will still suppress one side and amplify the other.

@PMOIndia @HMOIndia @narendramodi @AmitShah @MIB_India @AshwiniVaishnaw @MEAIndia @DrSJaishankar @FinMinIndia @nsitharaman @KirenRijiju

[Narendra Modi @narendramodi]

More strict actions against paper leaks to come in tomorrow’s Cabinet!

[Sumit @sumiiitkumar]

Nice sir , we support you ❤️

[Sumit @sumiiitkumar]

@i_pranavmehta After long time Pranav bhai

[Comman Man @CommanMan777589]

She woke up and chose not to restrain!!

Students sadak pe reh jaenge and Andolanjeevi ko naya task mil jaega 💀

[Aravind @aravind]

Nothing will happen. Remain calm & carry on. The state will take care. The escalation curve in India is generally very flat, the tolerance for anarchy very high, until the curve hits "subversion" on the x-axis. It then takes a 90 degree turn to "FAFO" on the y-axis.

[Aravind @aravind]

The DS toolkit - Steps for regime change:

1) Leverage a real grievance as the hook, amplify the discomfort even if it is resolved, then broaden it

Start with a concrete, emotional issue, affecting many. Amplify it and keep it alive even after it is addressed and resolved. Aggregate it with other pain points to incite more people. Create a wider audience.

2) Use social media reach, virality and memetic warfare as the accelerator

Launch with satire/memes to go viral fast (low barrier with Google form sign-up, humorous branding, social marketing). Create an identity to boost belonging. Incite emotional responses (anger, fear for the future). Use skilled digital strategists to create and push content.

3) Hybrid online-to-offline escalation and sustainement

Online outrage -> Small offline actions -> Gauge public response -> Time the offline escalation (months after initial issue and its resolution) -> Use online tools to gather support and bring more offline support at site. Install a well known public face as the leader of the protest. For future puppet leader needs as well.

4) Astroturfing and Amplification

True protestors with real concerns - few thousands. Astroturfing online farm accounts - millions. Simulate organic support, incite with new ideas, direct the protest using a cyber command, non-native cyber troops pushing narrative, and micro-targeted digital content. Use plants inside platforms and media to amplify an one sided narrative.

5) Political opportunism and piggybacking

Use country's politics to amplify the protests. Provide opposition with a platform to piggyback on to help galvanize more offline protestors. Incentivize them to use the opportunity to full extent showing them the possibility of a regime change.

6) Timing and goalpost-shifting for longevity

Build slowly or time actions after an issue peaks but while anger lingers. If the core problem is addressed, pivot to broader/political demands or "systemic change" to keep relevance. This prevents quick fizzling. Keep issue alive and galvanize more support.

7) Use the created protests and planted narrative to incite violence and blame it on the govt

Create a pre-emptive narrative of "govt going to crackdown" using force and violence. Plant a narrative of govt infiltrating protests. Orchestrate violence and blame it on the govt. Incite the general public and the world to condemn and get involved. Justify actions like storming the parliament, taking over govt institutions, and arson by protestors. Amplify protests multi-fold and make the regime buckle.

Saw all this executed in Ukraine, Sri Lanka, Bangladesh, and Nepal. Can India be regime changed like this? The clueless supremacist master of the DS and its dog who are trying this imagine so. What they don't understand is the dog is going to be destroyed and the master made to manage a crisis never seen before in the coming years.

[Office Of Vijay Patel @VijayGajeraO]

Highly Important.

I have been able to infiltrate a planned TOOLKIT, and it's extremely shocking to see the detailed planning and information about helpers.

I request you all to share it maximum.
Let's start to expose.👇

[Sajal Dhiman @SajalDhiman2]

“Kubernetes already has NetworkPolicies… so why does everyone keep talking about Cilium?” 🤔

Because Kubernetes defines the rules. It doesn’t enforce them.

NetworkPolicy is just an API object.

Without a CNI that implements it, it’s nothing more than YAML sitting in etcd.

This is where Cilium comes in.

✅ Enforces NetworkPolicies using eBPF
✅ Replaces kube-proxy for faster service routing
✅ Provides deep network visibility with Hubble
✅ Enables advanced security beyond basic L3/L4 policies
✅ Improves performance by avoiding iptables

Think of it this way:

Kubernetes says: “Frontend can’t talk directly to the database.”

Cilium says: “Got it. I’ll make sure that traffic is actually blocked.”

Writing a NetworkPolicy doesn’t secure your cluster.

Using a CNI that can enforce it does.

That’s why production Kubernetes teams care about Cilium.

#Kubernetes #DevOps #PlatformEngineering #CloudNative #Cilium #eBPF #K8s

[Uday👨‍💻 @uday_devops]

🚀 Top 5 Kubernetes Troubleshooting Scenarios

I put together a quick reference guide covering some of the most common Kubernetes issues you'll encounter across Dev, Stage, and Production environments.

Inside you'll find:

✅ Pod troubleshooting
✅ Service connectivity issues
✅ Node health checks
✅ Ingress debugging
✅ Persistent Volume (PV/PVC) problems

Plus essential kubectl commands and practical troubleshooting steps you can use in real-world environments.

Hope it helps fellow DevOps engineers and platform teams. 👇

[Tech Fusionist @techyoutbe]

https://t.co/JWXoeex9yt

[yourclouddude @yourclouddude]

S3 Storage Classes Simplified for you 🚀

[Uday👨‍💻 @uday_devops]

DevOps Interview Questions:

Linux:
• How do you create users and groups in Linux?
• Explain file permissions and ownership.
• What is `umask`?
• How do you troubleshoot high CPU, memory, and disk usage?
• Write a Bash script to monitor disk utilization.
• Difference between `cron` and `systemd` timers.

Git:
• Difference between Git Fetch, Pull, Merge, and Rebase.
• Explain Git Reset vs Git Revert.
• How do you resolve merge conflicts?
• Explain your branching strategy.
• How do you roll back a bad commit in production?

Docker:
• Explain the Docker architecture.
• Difference between Image, Container, and Volume.
• What are Multi-stage Builds?
• How do you optimize Docker image size?
• Explain Docker networking.
• How do you troubleshoot a container that keeps restarting?

Kubernetes:
• Explain Kubernetes Architecture.
• Difference between Deployment, StatefulSet, DaemonSet, and Job.
• Explain Services and Ingress.
• How do you troubleshoot CrashLoopBackOff?
• Difference between Requests and Limits.
• Explain ConfigMaps and Secrets.
• What are Liveness and Readiness Probes?
• Explain Node Affinity, Taints & Tolerations.
• How do you perform an EKS cluster upgrade?
• Explain HPA and Cluster Autoscaler.
• How do you back up an EKS cluster?
• Which Kubernetes monitoring and logging tools have you used?

Jenkins / CI-CD
• Explain your Jenkins pipeline.
• Difference between Declarative and Scripted Pipelines.
• How do you implement CI/CD for microservices?
• How do you secure Jenkins?
• What would you do if a pipeline suddenly starts failing?

AWS:
• Explain VPC Architecture.
• Difference between Security Groups and NACLs.
• Explain IAM Roles and Policies.
• Difference between EC2, Auto Scaling Group, and Load Balancer.
• Difference between EBS, EFS, and S3.
• How do private subnet instances access the internet?
• Which AWS services have you used in your project?

Terraform:
• Explain the Terraform lifecycle.
• What is Terraform State?
• Why do you use Remote Backend?
• Explain State Locking.
• What is Terraform Drift?
• Difference between Resource and Data Source.
• How do you manage secrets in Terraform?

Scenario-Based Questions:
• A Kubernetes Pod is in CrashLoopBackOff. How will you troubleshoot it?
• Jenkins pipeline is failing after a recent code change. What will you check?
• Terraform deployment failed halfway. How do you recover?
• Production application is not accessible after deployment. What steps will you follow?
• CPU utilization suddenly reaches 100% on production servers. How will you investigate?
• One Availability Zone goes down. How will your application continue serving users?

[Tech Fusionist @techyoutbe]

🚨 80% of production outages can be prevented with just a few design patterns.

If you're building distributed systems, these four resilience patterns should be part of your toolkit: Idempotency, Circuit Breaker, Transactional Outbox, and Saga.

Here's a simple visual that explains when and why to use each. 👇

[AsyncTrix @asynctrix]

☸️ Kubernetes in Plain English:

• Pod → the smallest running app unit
• Deployment → keeps your app running always
• Service → gives stable access to Pods
• Ingress → routes external traffic to apps
• Namespace → separates teams/environments
• ConfigMap → app configuration storage
• Secret → stores sensitive data securely
• Volume → persistent storage for containers
• Node → the machine running workloads
• Cluster → group of nodes managed together
• kube-apiserver → brain entry point of Kubernetes
• Scheduler → decides where Pods should run
• kubelet → worker agent on every node
• ReplicaSet → maintains desired Pod count
• HPA → auto-scales apps based on load
• DaemonSet → runs one Pod on every node

Master these and Kubernetes suddenly feels far less scary ☸️

[Sumit @sumiiitkumar]

@devops_nk Exactly

[Nandkishor @devops_nk]

GitHub Actions → Argo CD (CI/CD Flow)

Workflow Name
↓
Push / Pull Request
↓
Branch Filter
↓
Environment Variables
↓
Jobs (CI)
↓
Set Runner
↓
Checkout Code
↓
Install JDK
↓
Build with Maven
↓
SonarQube Scan
↓
Authenticate AWS Account
↓
Login to ECR
↓
Build Docker Image
↓
Trivy Security Scan
↓
Push Image to ECR
↓
Update deployment.yaml
↓
Commit Changes to Git
↓
Argo CD
↓
Sync with Git
↓
Deploy to Kubernetes

→ CI builds the image.
→ Git is the source of truth.
→ Argo CD handles deployment.

🎉 Congratulation app deployed to Kubernetes

Feel free to add more steps if I missed anything 👇

[Shubham Saboo @Saboo_Shubham_]

You can build, deploy and evaluate long-running AI agents using Claude Code, Codex or Antigravity.

Without writing a single line of code.

[Ayush @AyushSarode07]

Day 17 of building a distributed job scheduler in public 🚀

added dead-letter replay API + prometheus & grafana observability.

> added POST /dead-letters/{id}/replay API
> replays dead jobs as fresh QUEUED jobs
> added 9 Prometheus metrics (jobs, retries, DLQ, duration, workers)
> exposed /metrics on API, worker, and result collector
> built Grafana dashboard with throughput, retries, DLQ, latency (p50/p95/p99), and worker metrics.

Flow: Dead Letter -> Replay API -> New QUEUED Job -> Kafka -> Worker

next: idk lets see

[Skilled Coder @theskilledcoder]

System Design Uber : One Pager