Chronosphere’s Post

Your next #O11Y guide is here! In this chapter of the ongoing series, Chronosphere’s Eric D. Schabell looks into what makes #cloudnative #observability so crucial to your organization’s success: https://okt.to/QbmgjV

Your next #O11Y guide is here! In this chapter of the ongoing series, Chronosphere’s Eric D. Schabell looks into what makes #cloudnative #observability so crucial to your organization’s success: https://okt.to/cCBOpR

Your next #O11Y guide is here! In this chapter of the ongoing series, Chronosphere’s Eric D. Schabell looks into what makes #cloudnative #observability so crucial to your organization’s success: https://okt.to/X8PH9k

Your next #O11Y guide is here! In this chapter of the ongoing series, we look into what makes #cloudnative #observability so crucial to your organization’s success: https://okt.to/Do2gJj

Great introduction to the Observability Hierarchy of Needs by Balasubramanian Ramesh Babu @chronosphereio, just getting the series started so keep an eye on the blog for more installments soon! #cloudnative #observability https://lnkd.in/eVu4kenR

Discover how to transform your code by mastering observability strategy. Learn more in the latest addition to the #O11y guide: https://okt.to/mf4hpa

𝗢𝗽𝗲𝗻𝗧𝗲𝗹𝗲𝗺𝗲𝘁𝗿𝘆 𝗮𝗻𝗱 𝗖𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗙𝗲𝗲𝗱𝗯𝗮𝗰𝗸 - 𝗧𝗵𝗶𝗻𝗴𝘀 𝘆𝗼𝘂 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗸𝗻𝗼𝘄 𝗮𝗯𝗼𝘂𝘁 𝘆𝗼𝘂𝗿 𝗰𝗼𝗱𝗲 𝗮𝘁 𝗿𝘂𝗻𝘁𝗶𝗺𝗲! Tomorrow, catch Roni Dover's live online session here: https://lnkd.in/e3F78ez #continuousfeedback #opentelemetry

I'm happy to share my experience in deploying machine learning applications to production. 😃 In this post, I will discuss the approach I took. In the initial stage, I follow this simple approach 💡: ▪ I have started by setting up an EC2 instance and installing necessary packages. ▪ Cloned project repository onto the instance and ran the application there. ▪ Accessed the application via EC2 instance's public IP. Drawbacks ⚠ :  ▪ Project dependencies were tied to the instance, making it hard to manage.  ▪ Scaling and maintenance became challenging. Seeing the limitations of my initial approach, I adopted Docker for containerization. I leveraged containerization and CI/CD pipelines 🐳 : ▪ Dockerized the application for consistency across environments. ▪ Implemented CI/CD pipeline using GitHub Actions. ▪ Stored Docker images in ECR and used EC2 instances for deployment. ▪ Automated deployment process to reduce manual intervention. Drawbacks ⚠ :  ▪ Downtime occurred during container updates, affecting user accessibility.  ▪ Scaling still required manual intervention. Seeking to eliminate downtime, I explored the rolling update deployment strategy. Next, I follow Rolling Update Deployment Strategy 🔄: ▪ Adopted Rolling Update deployment strategy on Kubernetes (AWS EKS). ▪ Implemented replicas for high availability and resilience. ▪ Rolled out updates seamlessly without downtime. Drawbacks ⚠ :  ▪ Reverting to previous versions for troubleshooting was complex.  ▪ Managing multiple versions simultaneously was challenging. I encountered difficulties with version control and rollback capabilities. Finally, I adopted the blue/green deployment strategy 🔵🟢 : ▪ Embraced Blue/Green deployment strategy. ▪ Achieved zero downtime, smooth traffic switch, and quick rollback capability. ▪ Ensured consistent user experience with cost implications. Drawback ⚠ : ▪ Cost-expensive due to maintaining duplicate environments. ------------------------------------------------------------------------------------ Now tell me, How do you deploy your ML applications in production? Share your experiences in the comments below! Each approach taught me valuable lessons, but Blue/Green deployment emerged as the most reliable solution, offering seamless updates and minimal downtime. My project repo: https://lnkd.in/gQrdt4yi See my project demo showcasing a real-world Blue/Green deployment strategy in action https://lnkd.in/ejM2hDSU #deployment #kubernetes #ml #aws #awsdevops #githubactions #cicd #zerodowntime #clustermanager

O11y Guide, Cloud-Native Observability Pitfalls: Sneaky, Sprawling Mess #observability #o11y #cloudnative #guide https://lnkd.in/exYBr9fe

Meet Giant Swarm's Observability Platform with this blog post from Dominik Kress 🔍 Built on open source (Mimir, Loki, OpenTelemetry), it transforms system metrics into actual knowledge. Because observability should help you make better decisions, not just show pretty graphs. The cool part? Zero setup, built-in resource optimization, and it actually respects your team's expertise instead of prescribing how to use it.