Be DevOps Ready
0 likes•262 views
The document discusses DevOps, including what it is, why it is useful, and how to get ready to implement DevOps practices. Specifically, it defines DevOps as combining development and operations life cycles to reduce the time between coding and production. It recommends understanding concepts like continuous integration, continuous delivery, containers, infrastructure as code, source control, testing automation, and deployment strategies. The document provides tips on tools and skills to learn in order to be DevOps ready, and suggests ways to evaluate the success of DevOps implementations by measuring key performance indicators.
Be DevOps Ready
- 1. z DevOps Dr. Malinda Kapuruge 22 / 09 / 2021 ready … be …
- 2. z z Why DevOps? What is DevOps? How to get ready?
- 5. z z Once upon a time …
- 6. z Coding Testing Release Mgmt Operations time
- 7. z § A significant time gap between the ideation and production § Development and operations silos § Lack of insights § Low confidence to release § Increased cost of maintenance § Loss of revenue
- 8. z z Can we do better ?
- 9. z A typical week … Production System 24x7 up and Running New Features Bug Fixes Design Meetings Requirement Gathering and Refining Monitoring and Triaging
- 10. z Business Needs § Reduce time to market § Quickly fix production bugs § Without compromising quality § Better insights of production environment to make informed decisions
- 11. z What DevOps brings to the table? § DevOps is a culture change § Shared responsibility § Shared functionalities § Improved collaboration § Thinking end-to-end § Visibility end-to-end § Complements Agile (https://agilemanifesto.org/)
- 12. z Dev Ops Combines development and operations life cycles together …
- 15. z Understand CI and CD Infrastructure as Code Containers Source Control Scripting Language Test Automation Learn Linux Basics Deployment Strategies Monitoring Security Fundamentals Networking Fundamentals
- 16. z Understand CI and CD CI CD Source repo Automated tests Software package Artifact repo Infrastructure *Example only. Tools and steps may vary Versioning
- 17. z § Continuous Integration § Iterative small chunks of production ready code merged to a common branch § Automates tests as code is updated. Reject upon failure. § Continuous Delivery § Automated deployment - all the way to staging § Ready to be released at anytime § Continuous Deployment § Automated releases § Frequent (several times per day) § Thresholds, Alerts to find issues Continuous Integration Continuous Delivery Continuous Deployment Code build AUTOMATED AUTOMATED AUTOMATED Tests AUTOMATED AUTOMATED AUTOMATED Deploy to staging MANUAL AUTOMATED AUTOMATED Deploy to production MANUAL MANUAL AUTOMATED
- 18. z Infrastructure as Code § Databases, load balancers, servers and their provisioning mechanics § Machine readable definition files
- 19. z Infrastructure as Code § Multiple environments – same code § dev-*, tst-*, stg-*. prd-* § It worked on my machine dev § Time consuming § replicate bugs § recover § Destroy and redeploy § Cattle vs Pets dev tst stg prd
- 20. z Containers § Virtualization § Runtime issues § Same configuration - dev to prod § Repositories – store and reuse § Container orchestration – manage lifecycles § Small, portable, quick to spin § Container != VM Machine 1 Machine 2 Machine 3
- 21. z Containers
- 22. z Source Control § Both application code and infrastructure code § Controlled access § Auditability and recovery § Use conventions § Learn git basics ( https://rogerdudler.github.io/git-guide/ ) § Learn git best practices ( https://deepsource.io/blog/git-best-practices )
- 23. z Learn a Scripting Language § Automation scripts § Repetitive tasks § Consistency § Your own custom commands § Start with Python perhaps
- 24. z Test Automation § Test Pyramid § Tests run with every merge § Try to improve test feedback time § Pact tests (https://docs.pact.io/) § Detect integration issues early § Quality gates / code smells § Inspect code quality automatically Acceptance tests Integration tests Unit tests
- 25. z Learn Linux Basics § Shell commands § File system § SSH Source: www.phoenixnap.com/kb/linux-commands-cheat-sheet Source: www.ssh.com
- 26. z Deployment Strategies § Simple deployment. § Simply replace the instances with new version § OK for non-critical applications § Rolling deployment § Incrementally update instances in batches § Rollback if an error is detected § Blue – Green deployment § Two identical environments, live and staging § Test staging § Switch live and staging § Canary § Incrementally to a small subset of users § Targeted users Then Now 1.3.0 1.4.0
- 27. z Monitoring § Identify issues § Prevent failures § Usage patterns § Metrics and Alerts § System level as well as component level
- 28. z Security Fundamentals § Sensitive vs non-sensitive data § Principle of Least Privilege § Encryption methods – Symmetric vs Asymmetric encryption § Digital certificates - X.509 § Security standards, e.g., Oauth 2.0 § What’s an identity? § SaaS products
- 29. z Networking Fundamentals § Virtual Private Networks (VPN) § Disguise traffic § Cloud environments § e.g., AWS VPC § Public and Private Subnets § VPC links § Load balancers § Internet gateways § Bastion host / Jump Box
- 30. z
- 31. z Example: Deploy a microservice to ECS Fargate cluster in AWS
- 32. z Tips § Master one tool for each of the above categories. Learn more as you gain experience. § Document everything .Automate as much as possible. § It’s OK to be lazy smart § Use conventions, e.g., tst-1-customer-db § Externalize configuration, e.g., files, parameter stores § Logs ! § Do NOT re-invent the wheel § DevOps != SRE, DevOps != Security Specialist, DevOps != Agile
- 34. z KPI Description Change lead time From coding to deploy in production Change volume Features pushed to production per deploy Deployment frequency Number of deployments per day Success rate Successful deployments / total deployments Time to detect defects Mean time taken to detect a bug in the production Time to recover Mean time to recover Other application metrics For example, improvements in resource consumption due to devops feedback.
- 35. z z Thanks ! Time for a discussion…