Grand Treatise of Modern Instrumentation and Orchestration
summary
Practical recipe for Prometheus instrumentation in Go: use RED (request error duration) with histogram buckets and counters plus label vectors to get QPS, error rates and percentiles; expose the default /metrics HTTP handler to get runtime and process metrics for free; use gauges or function-backed metrics for channel length and utilization; implement a custom collector Describe/Collect and emit ConstMetric for exporters to scrape and convert third‑party servers (example: Apache server‑status) — demos compile and query correctly via PromQL.
AI-written summary · about
chapters (8)
- 0:00Intro
- 3:17Why Prometheus?
- 7:46Library-less exposition Just create the HTTP payload with your bare hands...
- 9:07Exposition with the client library
- 13:58I have metrics, now what? PromQL expressions for values at a point in time, graphs, and alerts
- 18:52Saturation of a buffered channel
- 20:34A simple exporter Example: Apache web server with mod_status
- 24:01How to collect multiple metrics from the result of a single function call? The Collector interface
related talks
22mPrometheus instrumentation: the Practical Way
47mGet Instrumented: How Prometheus Can Unify Your Metrics
31mHow to Export Prometheus Metrics from Just About Anything
40mInstrumenting GO Applications
36mMetrics - What's your code actually doing?
50mApplication metrics with Prometheus
30mIntro: Prometheus
35mThe RED Method: How To Instrument Your Services [B]