What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Nusku is an in-progress Linux continuous profiler that Ali Amer is building in Zig, with eBPF intended to observe running programs. Its goal is to help developers see what a process is doing as it runs—not only infer the cause of a slowdown from guesses or investigate after an incident. The design is ambitious, but Amer’s project introduction describes planned capabilities rather than a verified, finished profiler.
What Nusku is being built to observe
Amer describes Nusku as a way to inspect a running Linux program through several kinds of signals. The project introduction says it aims to show:
As an Amazon Associate I earn from qualifying purchases.
- CPU activity attributed to individual functions.
- Memory allocation locations, so a developer can investigate where allocations occur.
- Threads waiting on resources such as locks or I/O.
- Time spent in system calls.
- Network delays associated with packet retransmissions.
These are intended observations, not independently tested features. The introduction provides no benchmark, compatibility matrix, or evidence that every signal is implemented or available in a release. Amer also describes a remote workflow in which a machine streams a live view over a secure connection; that, too, is a design goal rather than a confirmed deployment capability. Amer’s project introduction
How the proposed architecture is supposed to work
The Agent comes first
The planned Agent would sit alongside a target process and collect observations through eBPF. Amer says a user should eventually be able to point Nusku at a running process or launch a process through it. Local observation and remote streaming are both described as intended workflows. The current development focus, however, is the Agent; Amer says the Controller is not the priority yet.
#1 Best Overall
A Controller is a later component
In the proposed design, a Controller would list running targets, let the user select one, and manage profiling sessions. That layer would matter when Nusku needs to watch multiple processes or support remote operation. It is part of the planned architecture, not a reason to assume those workflows are already usable.
Why Zig and eBPF?
Amer’s choice of Zig reflects an interest in systems programming, direct control over memory and allocation, and avoiding a garbage collector. He chose eBPF for kernel-level visibility without injecting code into the application being observed. These are the author’s design reasons; they do not establish that eBPF is the only suitable way to profile a program or that Nusku’s implementation has been validated across Linux environments.
Rank #2
How Nusku fits among Linux profilers
One concrete point of comparison is lightswitch. Its repository describes a Linux profiler with on-demand and continuous on-CPU profiling, a live terminal flamegraph, and saved profile formats. The repository identifies it as a Rust project using C for eBPF. That makes it a useful example of existing Linux profiling software, but it is not a head-to-head evaluation of lightswitch and Nusku.
Recommended Free Tools
| Decision point | Nusku, as described by Amer | lightswitch, as described in its repository |
|---|---|---|
| Signals | Aims to cover function-level CPU activity, allocation locations, waiting threads, syscall time, and network retransmission delays. | On-CPU profiling. |
| Operating mode | Local observation and remote streaming are intended workflows. | On-demand and continuous profiling are described. |
| Interface or output | A live view is part of the described goal. | Live terminal flamegraph and saved profile formats are described. |
| Implementation language | Zig, with eBPF for observation. | Rust with C for eBPF. |
The comparison is about stated scope, not equivalent maturity or performance. Nusku’s introduction does not establish its release status, implementation completeness, supported kernel versions, required privileges, or runtime overhead. Those details are necessary before a reader can judge setup requirements or use it as a production profiling choice.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.A project introduction, not a finished-tool review
Amer frames Nusku as a depth-first development project and a way to learn by building a systems tool. He also says there is no current plan to monetize it. Readers should treat the feature list and architecture as the author’s direction for the project, rather than as a promise of availability or a report of tested results. The introduction is dated October 4, but does not visibly establish a publication year.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
