Software Engineering Radio - the podcast for professional software developers SE Radio 706: Yechezkel "Chez" Rabinovich on Observability Tool Migration Techniques
28 snips
Feb 4, 2026 Yechezkel “Chez” Rabinovich, CTO and co-founder at Groundcover, a Kubernetes observability and migration specialist. He discusses why teams move observability stacks. He walks through discovery of hidden integrations, phased migration strategies, and validation techniques like dual-shipping. He also explains OpenTelemetry’s role and what it does not standardize, plus metrics to track migration health.
AI Snips
Chapters
Transcript
Episode notes
Hidden Inventory Of Monitors And Dashboards
- Yechezkel Rabinovich recalls customers arriving with hundreds to thousands of dashboards and monitors built over years by people who left the company.
- He describes a customer with a thousand unknown monitors who demanded a fully automated migration to avoid losing coverage.
Cost Shapes What You Observe
- Legacy observability often forces teams to reduce instrumentation because SaaS pricing and egress costs balloon.
- Groundcover's bring-your-own-cloud model keeps the data plane in the customer's account to avoid per-volume pricing and enable richer telemetry.
Vendor Incentives Shape Data Hygiene
- If an observability vendor makes money by ingest volume, incentives misalign and they may not help reduce noisy data.
- Pushing filters to node-level sensors avoids shipping and dropping spammy logs and reduces cost.