Dealing with Dark Debt: Lessons Learnt at
summary
Describes concrete practices at Goldman Sachs: a 25‑year SEC DB ecosystem with its own language slang, fortnightly releases, automatic pruning of unused code and tracked fix me comments; a centrally managed agent Java Runtime Monitor collecting ~150,000 samples/sec to map library usage and locate vulnerable callers; chaos/fault‑injection experiments and distributed tracing (Zipkin style) that pinpoint an incident where millions of new rows + stale stats trigger a full‑table scan. Culture: blameless postmortems, pair programming, SRE hackathons.
AI-written summary · about
chapters (29)
- 0:00Introduction
- 1:04Agenda
- 1:42About Goldman Sachs
- 2:29Engineering
- 3:00Footprint
- 3:42Environment Reliability
- 6:34Topics
- 7:06Prevention vs Cure
- 7:32What is SECDB
- 9:19What makes SECDB so successful
- 10:29linting and instrumenting
- 13:39production environment
- 14:50Java Runtime Monitor
- 17:36Blind Spots
- 19:57Debugging
- 20:26Tracing
- 21:00Tracing terminology
- 22:03Intermittent latency spikes
- 23:57Deep retry chain
- 24:22Culture
- 24:48Dos Donts
- 25:01Dont play whackamole
- 26:15Say no to dark matter developers
- 27:05Dont ignore your technical debt
- 27:43Increase runtime transparency
- 28:10Blameless postmortems
- 29:18Refactoring
- 29:50Recap
- 30:08Booth
related talks
46mManaging architectural debt in a hypergrowth context
52mEscaping Developer Nightmares
01h 35mRedesigning Legacy Systems - Keys to Success / Lessons Learned
01h 17mWhen Less Is More: Designing & Decoding Complex Distributed Systems
59mModernize Software and Tackle Tech Debt at Big Bank Scale
01h 03mIt's Abstractions All the Way Down...
54mResolving technical debts in software architecture
43mDeliver Fast with Confidence