3 Pitfalls Everyone Should Avoid with Cloud Data
summary
Cloud projects repeatedly blow budgets and ops because observability is massive and mismanaged: across customers about 60% useless telemetry arrives and most orgs store it for 13 month retention by default, a Hello World on a 4‑node cluster can generate ~0.5 TB of monitoring, and high cardinality labels explode costs. Real incidents include airport analytics failing due to egress fees and a load‑balancer‑of‑load‑balancers workaround. Adopt FinOps processes, require an exit strategy, and prune or aggregate before long term storage.
AI-written summary · about
chapters (7)
- 0:00Introduction to cloud data pitfalls
- 2:11Cloud data transfer costs
- 6:09Observability as data explained
- 15:46Developer stress and burnout in cloud ops
- 19:59FinOps as the crucial operations layer
- 23:06Summary and key takeaways
- 35:50Q&A session
related talks
59mBuilding the Best Cloud Data Architecture: Avoiding Costly Pitfalls
52mScaling Data Teams: Three Pitfalls and How To Avoid Them
36mFinOps tips for optimizing your cloud infrastructure and data costs
21mCloud-Native Observability for Financial Systems
26mFrom Billion...
25mData Demons: Unleashing Chaos or Driving Performance?
44mOpen Source Analytics, Public Clouds, and VC Mega-Investment
31mWhen the data levee breaks: Managing cognitive load in data-intensive projects