Lightning Talk: Who Are You? A Heuristic Approximation to Service Identification
summary
Proposes using Prometheus relabeling plus Kubernetes service discovery fields to identify apps when annotations aren’t possible, building extra scrape jobs that select pods by container name matching or port 9113 selection and then apply targeted relabel_configs to add or drop labels and metrics. Demonstrates keeping pods that match heuristics and adding a drop rule on the generic job to avoid double-scrapes. Notes limitations: heuristic false positives, no std annotation, and suggests a Prometheus option to prevent endpoints being scraped multiple times. Works but requires careful rule coverage.
AI-written summary · about
chapters (4)
related talks
11mLightning Talk: Grasp the Real Worth of Your Microservices by Brewing a Truth Poti...
02mLightning Talk: Ul To Debug Target Relabeling Rules
11mLightning Talk: Who Are You? Solving the Container ID Resolution With Cgroupv2
04mLightning Talk: Building Prometheus for Platform Engineering
42mMaking Resilient, Self-Adaptive, And Autonomous Kubernetes Clusters
31mEvolving Prometheus for More Use Cases
34mAutoscale your Kubernetes Workload with Prometheus - Frederic Branczyk, CoreOS
05mLightning Talk: Scrape Trolley Dilemma