Stability Patterns & Antipatterns
summary
Shows a real outage where idle DB connections get dropped by a firewall, TCP retransmits stall threads for 10 minutes and hang whole app servers until dead connection detection is enabled to send pings and restore availability. Demonstrates concrete fixes: configure DB pings or lightweight queries, add timeouts, circuit breakers and bulkheads, and run chaos testing with chaos monkey. Includes numbers, root cause packet traces, and the deployed fixes that actually stop cascading failures.
AI-written summary · about
chapters (35)
- 0:00Intro
- 2:53Availability
- 3:53Faults Errors
- 6:36Antipatterns
- 8:13Firewall
- 15:02wicked errors
- 16:50Debugging
- 20:58Microservice System Failure Mode
- 21:43Connection Pools
- 23:32Deep Links
- 24:41Marketing Calendar
- 25:17Scaling Effects
- 25:52PointtoPoint Messaging
- 26:59Shared Resources
- 28:11Traffic Patterns
- 30:54Slow Responses
- 32:00Excess Load
- 33:03Slow Response
- 33:20Un unbounded result sets
- 34:03Service architectures
- 34:15Testing
- 34:42Good News
- 35:56Circuit Breakers
- 37:13Circuit Breaker
- 38:55Sharing Circuit Breaker
- 39:45Bulkheads
- 41:02Bulkhead
- 41:28Fail Fast
- 42:02Applying Fail Fast
- 42:30Validation
- 42:55Decoupling
- 44:06Changing Middleware
- 44:30Different Architectural Styles
- 44:47Cracks in Metal
- 46:09Questions
related talks
34mAntics, Drift, and Chaos
32mHow does your application behaves when everything goes wrong?
34mDistributed Systems - What Can Go Wrong and How to Get It Right
01h 04mPredictably Unpredictable
41mFour advanced patterns to improve an application's resilience
56mThe Definition Of Insanity
16mFrom System TAYE7 to High Availability: What, Why, How?
48mFault Tolerance on the Cheap Making Systems That Probably Won't Fall Over