# Kakfa Summit - Monitoring Kafka Without Instrumentation Using eBPF ![rw-book-cover](https://readwise-assets.s3.amazonaws.com/static/images/article2.74d541386bbf.png) URL:: https://readwise.io/reader/document_raw_content/2509684 Author:: readwise.io ## Highlights > eBPF programs (user-defined, sandboxed bytecode executed by the kernel) allow > user-defined instrumentation on a live kernel image that can never crash, hang or > interfere with the kernel negatively ([View Highlight](https://read.readwise.io/read/01g86eg61kzn8mk2rzajpkf0k7)) > bpftrace > kubectl-trace ([View Highlight](https://read.readwise.io/read/01g86egzk5qvp8t64ck2fz906m)) > No instrumentation > ○ No code modifications > ○ No redeployments ([View Highlight](https://read.readwise.io/read/01g86ehefkb9mkx91fyawtktfz)) > Pixie’s approach > ● Pixie Edge Module (PEM) > deploys on every node > ● Capture data with eBPF > ● Process data in user-space > (protocol parsing) > ● Store data into tables for > querying by user ([View Highlight](https://read.readwise.io/read/01g86ehpvv6ghjgtqrf3ch5t23)) > Kafka observability is challenging ([View Highlight](https://read.readwise.io/read/01g86ejk6j5r7tj5gaa9nj8kxq)) --- Title: Kakfa Summit - Monitoring Kafka Without Instrumentation Using eBPF Author: readwise.io Tags: readwise, articles date: 2024-01-30 --- # Kakfa Summit - Monitoring Kafka Without Instrumentation Using eBPF ![rw-book-cover](https://readwise-assets.s3.amazonaws.com/static/images/article2.74d541386bbf.png) URL:: https://readwise.io/reader/document_raw_content/2509684 Author:: readwise.io ## AI-Generated Summary None ## Highlights > eBPF programs (user-defined, sandboxed bytecode executed by the kernel) allow > user-defined instrumentation on a live kernel image that can never crash, hang or > interfere with the kernel negatively ([View Highlight](https://read.readwise.io/read/01g86eg61kzn8mk2rzajpkf0k7)) > bpftrace > kubectl-trace ([View Highlight](https://read.readwise.io/read/01g86egzk5qvp8t64ck2fz906m)) > No instrumentation > ○ No code modifications > ○ No redeployments ([View Highlight](https://read.readwise.io/read/01g86ehefkb9mkx91fyawtktfz)) > Pixie’s approach > ● Pixie Edge Module (PEM) > deploys on every node > ● Capture data with eBPF > ● Process data in user-space > (protocol parsing) > ● Store data into tables for > querying by user ([View Highlight](https://read.readwise.io/read/01g86ehpvv6ghjgtqrf3ch5t23)) > Kafka observability is challenging ([View Highlight](https://read.readwise.io/read/01g86ejk6j5r7tj5gaa9nj8kxq))