brankopetric00's profile picture. DevOps Engineer | Grow your X account and get monetized @opentweetio

Branko

@brankopetric00

DevOps Engineer | Grow your X account and get monetized @opentweetio

Épinglé

I am super happy with the progress and features of @opentweetio If you want to generate X posts with AI, you can do that now within the OpenTweet itself. The rest is to pick the best posts and schedule them for posting.

brankopetric00's tweet image. I am super happy with the progress and features of @opentweetio

If you want to generate X posts with AI, you can do that now within the OpenTweet itself.

The rest is to pick the best posts and schedule them for posting.

Reduced Docker image size from 2.3GB to 180MB - Node.js application with massive image - Docker build taking 15+ minutes - Registry storage costs were ridiculous - Deployment rollouts painfully slow Image analysis: - Base image was full Ubuntu instead of Alpine - npm install…


Media upload test in production from OpenTweet

brankopetric00's tweet image. Media upload test in production from OpenTweet

Adding an API gateway to your monolith doesn't make it microservices architecture. You just added a very expensive proxy to your existing problems.


Switched from manual deployments to GitOps. What changed: - Deployments became predictable and traceable - Rollbacks went from 30 minutes to 30 seconds - Team confidence in releases increased dramatically - Debugging became easier with clear deployment history That transition…


Teaching someone to fish in DevOps means showing them how to debug, not just how to deploy.


Proper logging and monitoring is not only tools/infrastructure, but also the application code


Database was slow. Added read replicas, connection pooling, query optimization. Still slow. Turns out we were making 847 database calls per page load. Fixed with 3 lines of code to eager load relationships. Sometimes the solution isn't more infrastructure.


Stop running containers as root in production.


You're spending $50k/month on AWS because you provisioned everything in the most expensive regions with zero optimization. 'Cloud-first' isn't 'money-last.'


The best DevOps tool is a calm, experienced senior engineer who's seen this problem before. All the automation in the world can't replace someone who knows when NOT to automate something.


You can't be agile when your deployment process requires 7 approvals. Pick a lane.


Migrating from Jenkins to GitHub Actions because Jenkins is old while keeping the same pipeline logic. Congratulations, you moved your mess to the cloud.


Secrets in GitHub Actions, copied across 13 repos, rotated never. But at least they're not in the code anymore, right?


IAM policies with Action: '*' and Resource: '*' in production. Security team approved it because it was blocking the release. We deserve the breaches we get.


helm upgrade --install --force --wait --timeout 30m. You're not deploying anymore. You're negotiating with your cluster.


Your startup doesn't need Kubernetes. Your 5-person team managing 3 microservices doesn't need Kubernetes. Your monolith serving 1000 users doesn't need Kubernetes. But you deployed it anyway because: • 'Industry standard' • Resume padding • FOMO from conference talks Docker…


Microservices teams that share databases are playing distributed monolith theater. You split the code but kept the coupling. Now you have: - Database migrations blocking 6 teams - Schema changes requiring coordination calls - Transactions that span services - Data consistency…


SLOs without SLIs are just aspirational fiction. Your '99.9% uptime' SLO is meaningless when: - You measure uptime from your load balancer - Your database is down but health checks pass - Users can't log in but your homepage loads - Payment processing fails but you're…


Serverless isn't cheaper. It's vendor lock-in with a marketing budget. Lambda cold starts killing user experience API Gateway costs more than your compute DynamoDB bills that scale with your success Step Functions charging per state transition CloudWatch logs eating 30% of your…


Your AWS bill is 10x what it should be because you're running a data center in the cloud instead of using cloud services. EC2 instances that never shut down EBS volumes attached to nothing Load balancers for single instances RDS instances running at 5% CPU NAT gateways costing…


United States Tendances

Loading...

Something went wrong.


Something went wrong.