Enabling DevOps culture with a Platform Engineering team (TW)

DevOps is about creating a culture of shared responsibility in delivery teams. It is recommended to embed operations skills into delivery teams to reduce friction and enable them to deliver products to customers faster but whenever there is a need for significant investment in tooling and automation and teams lack the skills to self-manage a full application and operations stack, having a separate Platform engineering team can help. Organizations that consider establishing such a platform engineering team should be very cautious not to accidentally create a separate DevOps team.

In this presentation Adarsh Shah will talk about when and why Platform engineering team can be helpful. The team’s chief responsibility is to operate a platform which enables delivery teams to self-service deploy and operate systems with reduced lead time and complexity. The emphasis here is on self-service infrastructure and supporting tools, with delivery teams still responsible for supporting what they deploy onto the platform. He will also talk about challenges and lessons learned leading and being part of such teams. He will explain “The Three Ways: The Principles Underpinning DevOps” as mentioned in “DevOps Handbook” & “The Phoenix Project” and how it can be used to resolve some of these challenges faced by the team.

Transcript

1. Adarsh Shah Lead Consultant @ ThoughtWorks @shahadarsh shahadarsh@gmail.com Enabling DevOps culture with a Platform Engineering team Challenges & Lessons Learned
2. Sections • Different team structures for DevOps • Platform Engineering team • Challenges and lessons learned
3. Different team structures for DevOps
4. Dev & Ops
5. Dev Ops Software Development
6. Dev Ops Software Development IT buying Monitoring servers Respond to Outages Change Control Backup & DR planning Software Deployment Application monitoring & support Server provisioning & configuration
7. Dev & Ops Collaboration
8. Dev Ops Software Development IT buying Monitoring servers Respond to Outages Change Control Backup & DR planning Software Deployment Application monitoring & support Server provisioning & configuration
9. IT buying Monitoring servers Respond to Outages Change Control Backup & DR planning Software Deployment Application monitoring & support Server provisioning & configuration Software Development Dev Ops
10. Collaboration Software Deployment Application monitoring & support Server provisioning & configuration Dev Ops Software Development IT buying Respond to Outages Change Control Backup & DR planning Monitoring servers
11. Dev Ops Software Development Software Deployment Application monitoring & support Server provisioning & configuration IT buying Respond to Outages Change Control Backup & DR planning Monitoring servers
12. Advantages of this model • Improved communication & collaboration • Reduced feature lead time • Innovation • Agility • Software stability and quality
13. For this model to work • Dev and Ops - Shared vision • Organizational maturity • Need skills to self-manage operations stack
14. Team Structures for DevOps Dev Cross-functional team Platform Engineering team Platform <platform> Api driven monitoring tools analytics App Dev 1 Ops App Dev 2 server provisioning self service pipelines container platform Dev Ops Collaboration Dev & Ops collaboration Dev SRE HandOff - Meets Standards? Site Reliability Engineering (Google)
15. Platform Engineering team
16. Platform Engineering Team is responsible for operating a platform which enables delivery teams to self-service deploy and operate systems with reduced lead time and complexity. Platform Engineering team
17. Example responsibilities Platform <platform> Api driven monitoring tools analytics App Dev 1 Ops App Dev 2 server provisioning self service pipelines container platform
18. Example Workflow new microservice api call Platform Dockerfile <platform> Dev my-service.yml pipeline containers monitoring analytics server provisioning
19. Platform maturity Compute Data Network Containers Service Discovery Fault Tolerance Monitoring Build Pipelines Regression Testing Smoke Testing API Gateway Serverless Document Store Analytics Platform Maturity
21. App Dev 1 App Dev 3 Ops App Dev 2 App Dev 4 Platform Engg.
22. App Dev 1 App Dev 3 Ops App Dev 2 App Dev 4 Platform Engg.
23. App Dev 1 App Dev 3 Ops App Dev 2 App Dev 4 Platform Engg.
24. App Dev 1 App Dev 3 Ops App Dev 2 App Dev 4 Platform Engg.
25. Challenges & Lessons Learned with Platform Engineering teams
26. The Principles Underpinning DevOps The Three Ways Systems Thinking Amplify Feedback Loops Culture of Continual Experimentation & Learning
27. Avoid creating another silo
28. • Enablers not just Doer’s • Cross-team internships • Constant pairing with App dev teams • Lunch and learn sessions • Not a service center
29. Tackling Knowledge silos within the team
30. • Story pair rotation • Support pair rotation • Glossary of terms • Postmortem • Skills matrix exercise
31. Skills Matrix exercise Python Pipeline Kubernetes AWS Vault Consul Terraform David 3 2 1 2 2 1 1 Katie 1 1 2 3 2 3 1 Mark 3 3 2 1 3 2 2 Martha 3 3 2 2 1 3 3 Martin 3 3 2 3 3 2 3
32. Next sprint Python Pipeline Kubernetes AWS Vault Consul Terraform David 3 2 1 2 2 1 1 Katie 1 1 2 3 2 2 1 Mark 3 3 2 1 3 2 2 Martha 3 3 1 2 1 3 3 Martin 3 2 2 3 3 1 3
33. Building the right features
34. • Promote Product thinking • Delivery infrastructure as product • Observe and talk to customers • Key Performance Indicator driven • Product Owner Prioritization
35. Path to Production exercise Logging Monitoring Business Continuity Disaster Recovery Security Delivery Infrastructure Support Now Format AD integration Pipelines Soon Aggregation SLAs, KPIs Secrets mgmt Container platform Later Retention, Masking API for infra provisioning Near Release BCP test Failover test Pen testing Post Release
37. Being Agile
38. • Story Writing • Analysis is required but its technical analysis • Not Ticket based • Standup, Retros & other agile practices • Product Owner Prioritization
39. Embracing rapid adoption
40. • Showcases • Enablement • Maintainability • Build trust
41. Different models • Temporary vs Permanent
42. Summary • Different team structures for DevOps • Platform Engineering • Challenges • Avoid creating another silo • Building the right features • Being Agile • Embracing rapid adoption • Tackling knowledge silos within the team • Temporary vs Permanent
43. Questions Adarsh Shah Lead Consultant @ ThoughtWorks @shahadarsh shahadarsh@gmail.com