Probe: Program Traces as First Class State
summary
Library captures rich runtime state as maps (tags, timestamp, thread, dynamic bindings, auth, cause-id) instead of strings, pushes them off the hot path via core.async channels, filters/transforms subscriptions, and writes durable records to Cassandra (or an in-memory rolling buffer) for auditing and forensic queries. A probe macro records lexical context and the system can do runtime function instrumentation to log entry/exit and args. Deployed in a pre-production distributed healthcare stack on AWS, it materially improves tracing and auditing without blowing up latency.
AI-written summary · about
chapters (14)
- 0:00Start
- 0:27Overview
- 1:31Probe: Library plus Conventions
- 2:14Working with Distributed Systems
- 2:32Switchboard
- 4:44System Composition
- 6:19Actions and Operations
- 7:06Operations are Impure
- 7:49Need Instrumentation
- 8:18Auditing / Regulations
- 8:483. Online Model Updates
- 9:24Text Logs are Powerful
- 16:08Probe Dataflow
- 19:19Forensics Library
related talks
26mProfiling Mastery: Decode Your Data and Elevate Observability
25mPerfect Match: Correlating Continuous Profiling with Distributed Tracing for Stronger Observability
34mTrace-Profiling: A New Way About How to Track Application Behavior
45mTracing: A journey to tactical insights
53mSource Instrumentation for Monitoring C++ in Production
57mDistributed Tracing: How the Pros Debug Concurrent and Distributed Systems
40mContinuous Profiling, the missing piece in your observability puzzle!
43mGenerating and Analyzing Events