Ask HN: Continuous Profiling and System Metric
I have always been fond of continuous profiling for cpu and allocation and it seems there's been a major improvement in the landscape lately with a ebpf based profiling that is low overhead, easy to deploy and generally covers a lot (system level stuff for VM based language). It seems to be ebpf should also be good at capturing system metrics from cpu/memory utilization to more complicated one like IO/network latency, context switches and what not given the kernel level access.
It seems like a single ebpf based profiling tools should be able to capture almost anything performance related and present in a co-relatable manner. But while I'm finding tools for the profiling or some system monitoring stuff I'm not finding anything combining both. I wonder if I'm not doing a good search, it's not really possible as I imagine or as of now it's not been done because there's dedicated tool for metrics etc.
No comments yet.