Unikraft Performance Monitoring with Prometheus
summary
Describes porting a Prometheus exporter into Unikraft by adding it as a library running on a separate thread and using a ukstore intermediary (ProcFS-like) to expose metrics. Covers how ukstore provides build-time static and run-time dynamic entries for libraries to publish memory, networking and thread-locking metrics (with filesystem/CPU in progress), and how a low-dependency exporter was patched to collect those metrics with minimal performance impact.
AI-written summary · about
related talks
34mMonitor My Socks: Using Prometheus in a Polyglot Open Source Microservices
32mDemocratizing Application Monitoring with Prometheus
24mPractical Kubernetes Monitoring with Prometheus
37mMonitoring Kubernetes with Prometheus
34mAutoscale your Kubernetes Workload with Prometheus - Frederic Branczyk, CoreOS
31mHow to Export Prometheus Metrics from Just About Anything
32mImproving Kafka monitoring with native Prometheu...
40mThe Prometheus monitoring system