Apps, Stacks, and Frameworks
summary
A team replaces relational hacks with MongoDB to match a JSON model, deploys a five‑node cluster, and ships quickly; it works initially but three months later CPU and disk thrash after users generate 15 million documents because queries and indexes are wrong, no load testing ran and SOPs are ignored. The talk gives concrete failures—misused schema, poor indexing, no load testing, and a rushed set‑it‑and‑forget‑it deployment—so judge by these specifics.
AI-written summary · about
chapters (25)
- 0:00Start
- 1:32Developer Advocate
- 5:26DEVOPS Strong
- 6:03New Features Web UI
- 7:58Team Agrees on Solution
- 8:45MongoDB Infrastructure
- 9:27Deploy new Build to Production
- 9:58Project Goals & Deadlines Met
- 10:09Happy Customers
- 10:29Alerts from MongoDB Clusters
- 10:42Performance Declined & Failures
- 11:02Findings
- 11:22Poorly Written Queries & Bad Indexing
- 11:57Triage Move Records
- 12:50Bad Programming & Queries
- 13:29Improper Vetting
- 13:59Avoiding Shiny Objects Syndrome
- 14:16Don't Believe the Hype
- 16:48Don't Break Protocol
- 17:22Set and Forget It
- 17:35Technology is about people
- 19:31Mentor
- 19:57Make Hard Choices
- 20:48Something's broke? Fix It
- 22:48Communicate
related talks
29mApps, Stacks, and Frameworks: Avoiding “Shiny Object” Syndrome
36mApps, stacks & frameworks: Avoiding "Shiny Object" Syndrome
29mThe Front-end Takeover
12mBuild The Next Big Thing With MongoDB’s Developer Data Platform
20mtech organizations are not agile...at all with Daniel France and Jaden Lemmon
27mFacing Frontend's Existential Crisis
12mScaling a Monolith, Is A Four-Letter Word
26mArchitecting for scale