working on kepler and understanding how it used to work with the model estimators.
learning about cpu/[0,N]/msr, rapl and how does kepler use that to estimate.
wrote a blog post on resolve-virtmanager-network
able to make zero downtime deployments when there is a event based worker processing events and about draining it before making releases, all thanks to helm
learning how to better check for suitable limits, requests for cpu and memory for k8s workloads.
learning about placement groups usage any Cloud provider also the microsecond latencies and how to measure them.
learning about how to instrument a go app with opentelemetry for better logging and tracing for a production workload having multiple external services and find bottlenecks.
learning about longhorn RWX across workernodes and how it does replication among them and what happens if all the nodes loose data when its being used by applications.
learning about Linux Bluetooth and Audio codecs and AAC vs HAAC… and things ike pipewire and bluez and alsa and pulseaudio and how they interact with each other.
k8s sidecar containers which is now official feature since k8s 1.30