🎉 Just crossed 100+ followers on X!
It might be a small milestone, but it's still worth celebrating. Every follow, like, repost and interaction means a lot. 🙌
A big THANK YOU to everyone who has followed and supported me along the way. I genuinely appreciate it!
I'm here to learn, share, grow, and connect with like-minded people.
If you're into Tech, DevOps, Cloud, AI, or anything in between, let's connect! 🚀
Here's to learning and growing together! 🤝
#TechCommunity #DevOps #CloudComputing #AI
Two merges, a minute apart. Two deploys. The older one finishes last and rolls prod back.
One block in GitHub Actions stops it:
concurrency:
group: deploy-prod
cancel-in-progress: false
One deploy at a time, newest queued run wins.
How do you stop deploys racing?
Before you run anything against AWS, check who you are:
aws sts get-caller-identity
It shows the account and role your CLI is using right now. One look saves you from running a cleanup script against prod with the wrong profile.
What's your guard against the wrong account?
Every intimidating AWS architecture diagram started as one EC2 instance.
Here's what happens in between, one problem at a time:
1. One EC2 instance
App and database on the same box. Cheap and simple, and one failure takes everything down.
2. Database moves to RDS
Backups and patching are managed now, and the app server becomes replaceable.
3. Load balancer plus a second instance in another AZ
Losing one server, or one AZ, is no longer an outage.
4. Auto scaling and a cache
ElastiCache takes repeated reads off the database. Traffic spikes stop being incidents.
5. Static files to S3, behind CloudFront
Images and scripts load from close to the user, and your servers stop serving them.
6. Queues and events
Slow work goes to SQS and workers. Services talk through events (EventBridge, Lambda).
Don't start at step 6. Each step solves a problem you should actually have first.
bind: address already in use
Don't reboot. Ask who has the port:
sudo ss -ltnp 'sport = :8080'
You get the process name and PID. Nine times out of ten it's an old copy of your own app that never shut down.
ss or lsof -i: which one is muscle memory for you?
Every intimidating AWS architecture diagram started as one EC2 instance.
Here's what happens in between, one problem at a time:
1. One EC2 instance
App and database on the same box. Cheap and simple, and one failure takes everything down.
2. Database moves to RDS
Backups and patching are managed now, and the app server becomes replaceable.
3. Load balancer plus a second instance in another AZ
Losing one server, or one AZ, is no longer an outage.
4. Auto scaling and a cache
ElastiCache takes repeated reads off the database. Traffic spikes stop being incidents.
5. Static files to S3, behind CloudFront
Images and scripts load from close to the user, and your servers stop serving them.
6. Queues and events
Slow work goes to SQS and workers. Services talk through events (EventBridge, Lambda).
Don't start at step 6. Each step solves a problem you should actually have first.
@twtayaan Even now it's the GOAT as it doesn't have vendor lock-in and helps in no-hassle migration in my opinion 😅
Just https://t.co/ns3tZNzbpg irrespective of the tools you are using!
Absolutely! Build and operate a real DevOps environment.
Linux: Provision linux servers, manage users and permissions, configure SSH, write shell scripts, monitor processes in the server, and troubleshoot any failures.
Networking: Configure VPCs, subnets, routing tables, security groups, DNS, load balancers, and troubleshoot connectivity between services.
Databases: Deploy PostgreSQL, manage users and permissions, automate backups, configure replication, monitor performance, and practice disaster recovery.
Put it all together: Build a production like environment where you build an application into Docker images, automate infrastructure with Terraform, set up CI/CD with Jenkins or GitHub Actions, Deploy the application to Kubernetes cluster and monitor everything with Prometheus and Grafana.
Intentionally break things, troubleshoot incidents, automate repetitive tasks, and document your fixes.
That's how you go from knowing DevOps tools to actually thinking and operating like a DevOps engineer.
Great post reflecting the happenings on-ground!
Traefik’s kubernetesIngressNGINX provider offers a low-disruption migration path from the retired ingress-nginx controller.
It supports common NGINX annotations, allowing most existing Kubernetes Ingress manifests to work with minimal changes.
Teams can migrate progressively, validate compatibility, and later adopt Traefik IngressRoute or Gateway API.
Key takeaway: Avoid a big-bang migration while improving security and moving to an actively maintained ingress solution. Test annotation compatibility, especially authentication and rewrites, before cutover.
Docs: https://t.co/8Fpo1A1VPp
Delete a Pod and its replacement gets a new IP. That's why Services exist.
A Pod is one or more containers sharing one IP, localhost and volumes, on one node, and it's disposable by design.
Kubernetes from zero, part 2:
https://t.co/5I4jujTwsA
🚨 Your Kubernetes cluster might be running perfectly, while quietly exposing itself to attackers.
That's where Kubescape comes in. 🔐
An open-source Kubernetes security tool that helps you discover what you might be missing:
🔍 Detect security misconfigurations
🛡️ Identify risky RBAC permissions
📋 Check CIS Kubernetes benchmarks
🐳 Scan container images for vulnerabilities
⚠️ Uncover potential security risks before attackers do.
The best part? You can scan your cluster instead of manually hunting through hundreds of YAML files.
If you work with Kubernetes, this is definitely worth exploring.
⭐ GitHub: https://t.co/VcYxSg24s0
Secure your cluster before someone else finds the weakness.
Thanks for detailed information! I would add little additional info
Exit 137 + OOMKilled means the kernel SIGKILLed the container for exceeding its memory limit, confirm with `docker stats` + `docker inspect`, and set explicit memory limits to prevent it. Filter `docker events` (`--filter event=die`) and pair with recent logs to catch restart loops quickly.
🎉 Just crossed 100+ followers on X!
It might be a small milestone, but it's still worth celebrating. Every follow, like, repost and interaction means a lot. 🙌
A big THANK YOU to everyone who has followed and supported me along the way. I genuinely appreciate it!
I'm here to learn, share, grow, and connect with like-minded people.
If you're into Tech, DevOps, Cloud, AI, or anything in between, let's connect! 🚀
Here's to learning and growing together! 🤝
#TechCommunity #DevOps #CloudComputing #AI