Serg Hospodarets
Engineering and technology leader, speaker, Web and Cloud infrastructure enthusiast.
Title | Managed team size |
---|---|
Developer, Tech. Leader | ~10 |
Architect, Eng. Manager | ~20 |
Director of Engineering | 50+ |
VP, Global Head of Eng. | 100+ |
Organization | Main request |
---|---|
💰 Finance (CFO) | Cost efficiency |
Product (CPO) | ⏰ Time to market |
🔐 Security (CIO) | - security checks - fast security fixes |
Leaders, CPeopleO | - more common and modern tech stack |
COO | 📊 Be ready for M&A |
Board, CEO | Fast results and future projection |
🏛 Architects | Common architecture |
Eng. and Operations | Less cognitive load |
- up to 20 days of downtime per year
- 1 day MTTR
- 124 different Vendors, 30% of infra cost reduction hard request
- Ad hoc and per project security checks and fixes rollout
- Prod deployments once per a month
1. Way to apply DevOps principles in practice in large orgs.
2. Helps to enable Team Topologies
First step- simple static docs:
- Docusaurus- create simple versioned doc sites
- Slate- simple static docs for API
Next step:
- Backstage- open-source Internal Development Portal
For scale:
Typical team structure
Development
Platform
overview
- Storybook - Frontend workshop for UI development (React, Vue, Angular, JS)
- Create React apps self-service - Provide own app templates
- Customize app config - own app configs
Create a microservice [optionally, with microfrontend]
- https://www.jhipster.tech/creating-an-app/ (CLI)
- https://start.jhipster.tech/generate-application (Web-based)
Other:
- Provide own application templates in Jhipster
- Create own app templates in Backstage
Typical team structure
Development
Platform
overview
- Provide common artifacts and container solutions
- Environments
- Common Observability
- Upgradability
- Introduce common Security and Quality checks
etc.
Typical team structure
Development
Platform
overview
- Use for automated rollout of new features and versions across Platform customers
Front-End: Codeshift / Codemods
Microservices, apps and infra: JHipster Upgrading application
Dependencies: Renovate- multi-platform and multi-language.
Area | Outcome |
---|---|
Platform-based tech. | 50+ apps 100+ microservices, 20+ microfrontends |
📝 DORA metrics | MTTR decreased from 1 day to 4 hours Availability went from 95% (20 days down) to 99.999% (5 min/year) |
⏰ Time to market | Once per month prod deployments ➡️ to daily deployments and weekly releases. |
🔐 Security checks - fast security fixes |
Applied Common Platform security guardrails (shared CI/CD) Supply chain fixes delivered in ~3 hours to prod across portfolio |
Other Metrics | Time to onboarding decreased 2.3 times eNPS score improved for +12 |
- Pilot team(s) is super important to become evangelists
- Treat Platform as a product:
1. Internal teams are your customers: provide support, Slack etc. channels
2. Invest in Platform PM and DevX culture
3. Evangelise- regular releases, demos, newsletters
- For UI adoption put heads together with the UX, infra- DevOps, security- SecOps
- Cooperate with Architects/Directors/VP/CXO across the org
- Jackpot Platform strategy- find a way and deliver your platform to your company customers
Inspirational 🗣
- Prefer golden path over golden cage
- Foster the Pull vs Push culture
- Work with leadership for a top-down support for bottom-up enablement
ㅤ
By Serg Hospodarets
Do you want to move to the next stage of DevOps- Platform Engineering? Then this talk is for you- from discussing Why, to actually How to create platform engineering organisation + demos which tooling and approaches to apply to create the team, the Platform, culture, and the proper process. Resources and Youtube video: https://hospodarets.com/platform-engineering-talk
Engineering and technology leader, speaker, Web and Cloud infrastructure enthusiast.