48.5% of organizations are already using OpenTelemetry. Another 25.3% want to implement it but are stuck—blocked by the biggest adoption barrier: instrumenting existing applications requires code changes, rebuilds, and coordination across every team. In November 2025, OpenTelemetry released an answer: eBPF Instrumentation (OBI), which instruments every application in your cluster—Go, Java, Python, Node.js, Ruby—without touching a single line of code. Here's how to deploy it in production, what it can and can't do, and when you still need SDK instrumentation.
Your team spent nine months implementing Backstage. The portal looks beautiful. You have 47 services cataloged. Internal adoption? Eight percent. Three developers use it regularly. The rest go directly to AWS console, kubectl, and GitHub.
Sound familiar? You're not alone. Backstage's own community acknowledges the 10% adoption problem—average adoption stalls at 10% within organizations, and teams require 7-15 FTE to maintain it. Here's why it happens, and the decision framework I wish we'd had before investing 2 person-years of engineering time.
🎙️ Listen to the podcast episode: Backstage in Production: The 10% Adoption Problem - Jordan and Alex discuss the real costs, alternatives, and decision frameworks for choosing Backstage vs Port, Cortex, and custom portals.
The data is in, and it's turning conventional wisdom on its head.
If you're grinding through cloud certifications thinking it's your ticket to the top salary brackets, I've got news that might make you reconsider your entire learning roadmap.
Your platform engineering team manages 130+ tools. Your engineers use 10-20% of their capabilities. You're spending $400k on AI tools that 71% of your developers don't trust.
Welcome to platform engineering economics in 2025—where the hidden costs are killing your ROI, and traditional metrics aren't telling the real story.
Tool sprawl is costing you more than licenses: With enterprises managing 130+ tools, engineers lose 3.8 hours daily to context switching (23 min per switch), custom tools consume 20-30% of team capacity for maintenance, and companies spend $400k on AI tools with only 29% developer trust. The real ROI metrics that matter: 40% fewer outages, 60% faster incident recovery, and understanding that downtime costs $500k-$1M per hour. Consolidate tools, measure outcomes not outputs, and treat your platform as a product developers actually want to use.
The Tool Sprawl Crisis Nobody Wants to Talk About
Let's start with a number that should make every CTO pause: Engineers are managing an average of 16 monitoring tools. When SLAs get strict? That number jumps to 40.
As one frustrated platform engineer put it: "Teams use only 10-20% of tool capabilities but still pay full price."
The scale varies, but the problem doesn't:
Small companies: 15-20 tools
Medium businesses: 50-60 tools
Large enterprises: 130+ tools
And here's the kicker—global spend on security tools alone is projected to hit $261 billion by 2025. That's billion with a 'B'.
💡 Key Takeaway: Tool sprawl isn't just about license costs. With 16 monitoring tools on average (40 for strict SLAs), enterprises managing 130+ tools are paying for features they barely use. Teams utilize only 10-20% of tool capabilities while paying full price—that's like buying a sports car to drive to the grocery store.
