From 56d0ddb25ce815533faab4c771b31eaee2564411 Mon Sep 17 00:00:00 2001 From: Ivan Babrou Date: Sat, 18 Jan 2025 22:34:14 -0800 Subject: [PATCH 1/2] Fix remaining 100Hz -> 19Hz mentions --- docs/ingestion.md | 2 +- docs/parca-agent.md | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/ingestion.md b/docs/ingestion.md index c8d16d8..707c5f5 100644 --- a/docs/ingestion.md +++ b/docs/ingestion.md @@ -4,7 +4,7 @@ There are two ways to ingest data into Parca. ## Push based -The recommended way to use for generic profiling is Parca Agent. Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 100 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. +The recommended way to use for generic profiling is Parca Agent. Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. The collected data can be [sent to a Parca server](https://buf.build/parca-dev/parca/docs/main/parca.profilestore.v1alpha1), where it can be queried and analyzed over time. diff --git a/docs/parca-agent.md b/docs/parca-agent.md index a33bb7d..7a15fb5 100644 --- a/docs/parca-agent.md +++ b/docs/parca-agent.md @@ -1,6 +1,6 @@ # Overview -Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 100 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. +Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. The collected data can be viewed locally via HTTP endpoints and optionally be configured to be sent to a Parca server, where it can be queried and analyzed over time. From d5fc305aa0ba3c97e36ea40fc3db758af04ea06d Mon Sep 17 00:00:00 2001 From: Ivan Babrou Date: Sat, 25 Jan 2025 16:05:20 -0800 Subject: [PATCH 2/2] Clarify that 19Hz is per logical CPU --- docs/ingestion.md | 2 +- docs/parca-agent-design.md | 2 +- docs/parca-agent.md | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/docs/ingestion.md b/docs/ingestion.md index 707c5f5..b4b9528 100644 --- a/docs/ingestion.md +++ b/docs/ingestion.md @@ -4,7 +4,7 @@ There are two ways to ingest data into Parca. ## Push based -The recommended way to use for generic profiling is Parca Agent. Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. +The recommended way to use for generic profiling is Parca Agent. Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second per logical CPU and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. The collected data can be [sent to a Parca server](https://buf.build/parca-dev/parca/docs/main/parca.profilestore.v1alpha1), where it can be queried and analyzed over time. diff --git a/docs/parca-agent-design.md b/docs/parca-agent-design.md index 652b7dd..24c1f82 100644 --- a/docs/parca-agent-design.md +++ b/docs/parca-agent-design.md @@ -1,6 +1,6 @@ # Design -Parca Agent implements a sampling profiler, to sample stack traces [19 times per second](#cpu-sampling-frequency) via eBPF. It tracks user space as well as kernel-space stack traces. From the raw data, it builds a [pprof](https://github.com/google/pprof) formatted profile and optionally sends it to a Parca server where it is stored and can be queried and analyzed over time. +Parca Agent implements a sampling profiler, to sample stack traces [19 times per second](#cpu-sampling-frequency) per logical CPU via eBPF. It tracks user space as well as kernel-space stack traces. From the raw data, it builds a [pprof](https://github.com/google/pprof) formatted profile and optionally sends it to a Parca server where it is stored and can be queried and analyzed over time. Parca Agent is a whole-system profiler. It collects stack traces from all the processes that run on the host system. This provides more insights about all the aspects of the system to the user. Please see our [blog post](https://www.polarsignals.com/blog/posts/2022/08/24/system-wide-profiling/) about internals of this mechanism. diff --git a/docs/parca-agent.md b/docs/parca-agent.md index 7a15fb5..259a70d 100644 --- a/docs/parca-agent.md +++ b/docs/parca-agent.md @@ -1,6 +1,6 @@ # Overview -Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. +Parca Agent is an always-on sampling profiler that uses [eBPF](https://ebpf.io/) to capture raw profiling data with very low overhead. It observes user-space, and kernel-space stack traces 19 times per second per logical CPU and builds [pprof](https://github.com/google/pprof) formatted profiles from the extracted data. For in-depth detail and explanation refer to the the [Parca Agent Design](/docs/parca-agent-design) documentation. The collected data can be viewed locally via HTTP endpoints and optionally be configured to be sent to a Parca server, where it can be queried and analyzed over time.