DevOps is for startups!Enterprises do platform engineering and here is why - Jürgen Süßner
summary
Enterprise with ~8,500 employees and ~2,000 developers builds a self-service platform that uses Kubernetes clusters as a self-service, Cloud Foundry, Flux, Concourse, Prometheus/Grafana, Kafka/RabbitMQ and Postgres/Mongo/Redis to push day‑2 automation. Platform teams (~10 core, ~50 product teams) offer database-as-a-service and monitoring-as-a-service with API booking, internal chargeback, automated zero‑downtime upgrades and Concourse pipelines that pull fixes -> staging -> dev. Result: developers regain up to four productive days a week versus heavy ops toil, and the platform is treated and measured as a product.
AI-written summary · about
chapters (26)
- 0:00Introduction and speaker background
- 1:27From simple app to complex infrastructure
- 8:27Challenges of managing multiple clusters
- 10:29Developer productivity decline in enterprises
- 11:10Platform engineering as a solution
- 13:13Self-service platform offerings
- 17:26Database and messaging as services
- 18:50Monitoring as a centralized service
- 20:13Pre-baking security and compliance
- 21:17API gateways and authentication flows
- 22:00Pipeline management and ownership
- 23:32Application templating and best practices
- 24:36Community-driven platform development
- 26:01Measuring productivity and KPIs
- 27:26Mindset shift and cultural change
- 29:36Summary and key takeaways
- 31:05Q&A session begins
- 31:28Limiting developer cluster growth
- 32:33Team size for platform management
- 32:54Cluster configuration rationale
- 33:37Balancing developer freedom and constraints
- 34:39Legacy workload migration
- 36:05ITIL vs product-based platform model
- 37:11Platforms and cost reduction
- 38:14Reducing ops burden for developers
- 38:35Self-service tools and future plans
related talks
28mDevOps meets reality
56mStill Doing DEVOPS? Platform Engineering Is the Next Big Thing
41mDevOps for Java Developers (or maybe against them?!)
37mDevOps, 12 Factor and Platform Engineering
10mDevOps in the Enterprise
37mDevOps for Java Developers (Or Maybe Against Them?!)
44mDevOps for developers (or maybe against them?!)
31mDevOps for Java Developers (Or Maybe Against Them?!)
