3 Ways You’re Screwing up Platform Engineering – And How to Fix It • Steve Smith • GOTO 2024
This YouTube video by Steve Smith discusses three common ways platform engineering projects fail, emphasizing the crucial need for a product mindset. Key points include:
I. Three Common Platform Engineering Screw-Ups:
-
Power Tools: Over-reliance on heavyweight tools (Kubernetes, Kafka, Istio) leads to high platform costs and significant unplanned work for delivery teams. The solution is to lift and shift workloads from heavyweight to lightweight alternatives, prioritizing the capabilities causing the most pain. Measure internal customer value, cost, and platform cost to track progress.
-
Technology Anarchy: Giving delivery teams complete autonomy without sufficient technical alignment results in inconsistent tech stacks, increased complexity, and reduced overall velocity. The solution is to achieve “aligned autonomy” by establishing clear guidelines, expectations, and consequences for technical choices, essentially creating “paved roads” for development. This is visualized with a 2x2 matrix showing the balance between autonomy and alignment.
-
Ticketing Hell: Treating the platform team as an operational service desk, relying heavily on tickets, creates massive backlogs, slow turnaround times, and burnout. The solution is to automate high-demand tasks, creating self-service capabilities and reducing the platform team’s workload. Focus on reducing lead time, not just task completion time.
II. The Underlying Problem: Lack of Product Mindset
The core issue behind all three screw-ups is the absence of a product mindset. Platform teams should act as enablers, not gatekeepers, focusing on delivering value to internal customers (delivery teams) and iteratively improving the platform as a product. This includes:
- Treating the platform as a product: Not a project with a defined end.
- Empowering the platform team: Enabling experimentation and risk-taking.
- Prioritizing metrics: Internal customer value, cost, and platform cost.
- Collaboration and communication: Regularly engaging with delivery teams.
- Automation and self-service: Shifting tasks from the platform team to delivery teams wherever possible.
III. Anecdotal Evidence
The speaker uses numerous anecdotes from his experience and clients to illustrate the problems and their impact, highlighting the negative consequences of not adopting a product-focused approach.
In essence, the video argues that successful platform engineering requires a shift from a project-based, operational approach to a product-based, customer-centric approach, prioritizing speed, reliability, and collaboration.