How simplifying software can save your engineering teams' time
summary
A messy ETL and fleet of redshift clusters gets replaced by an event-based pipeline that writes into elasticsearch, trading custom aggregation-heavy queries for a known toolset. Build effort drops from an expected six months to six weeks, operating cost falls from about $70k a month to roughly $5k, many more engineers ramp quickly, and the feature ships with far fewer outages and on-call pages.
AI-written summary · about
chapters (17)
- 0:00Intro
- 1:15What is respond insights
- 4:10Understanding the problem
- 4:25Answering questions
- 4:39Understand the problem
- 5:33Choosing boring technologies
- 5:45Donald Trump quote
- 6:26Five questions
- 7:07Can I reuse existing technology
- 7:36I already had answers
- 8:51Quote
- 9:15Tactics
- 10:35Mappings
- 11:02Cost
- 11:50Boring
- 12:00Boring gets a bad rap
- 12:10Outro
related talks
38mData Secrets from a Platform Engineer
26mMore Culture, More Engineering, Less Duct Tape v3.0
20mAutomatise everyday work
35mStreamlining Development Processes through Self-Service Optimizations
30mHow we increased developer productivity 2x
28mMoving to 1ES
07mFrom Scaling APIs to Scaling Data to Scaling Your Team
22mBuild Smarter, Don't Just Build