Saved by In-Memory NoSQL
summary
Claim: use in-memory NoSQL as a real-time TCP logging endpoint to capture encoded debug logs (short numeric codes from a mapping table) so every state, event and transition becomes a compact, persistent black-box record. Implementation: per-server in-memory store with a background batch job that flushes to disk/DB, plus a decoder + visualization pipeline to render activity diagrams. Result: after data-structure tuning this meets acceptable latency but needs careful engineering and has memory-size caveats.
AI-written summary · about
chapters (14)
- 0:00Intro
- 1:04Airplane Black Boxes are
- 2:17Black Box Architecture
- 4:23The State Machine - Miniature Version
- 4:58Example
- 5:37The encoded Logging - Create a mapping Table
- 5:53The log files will look like
- 6:37Black Box Debugging - Things we need at our end
- 8:00Some Logging Strategies
- 8:26To be able to behave like Black Box
- 8:59The Architecture
- 9:41Second Try - SQL Database Write
- 10:59After Third Try - NoSQL In-Memory Database Write
- 12:27NoSQL In-Memory Database write - Further Optimization
related talks
07mthe state of in-memory
44mJava In-Memory Database-Apps with MicroStream
01h 01mMemory First An introduction to Redis
51mIn-Memory Computing - The Big Picture
46mEdge Performance with In memory NoSQL
48mIn-Memory-Computing - The Big Picture
50mHow to Choose an In-Memory NoSQL Database Wisely
16mScaleDB: A Scalable, Asynchronous In-Memory Database