zap
zigzap/zap
blazingly fast backends in zig
zigzap/zap
blazingly fast backends in zig
Zap is the zig replacement for the REST APIs I used to write in python with Flask and mongodb, etc. It can be considered to be a microframework for web applications.
What I needed as a replacement was a blazingly fast and robust HTTP server that I could use with Zig, and I chose to wrap the superb evented networking C library facil.io. Zap wraps and patches facil.io - the C web application framework.
After having used ZAP in production for a year, I can confidently assert that it proved to be:
Exactly the goals I set out to achieve!
zig-master
branch. An example of how to use it is
here. Please note that the
zig-master branch is not the official master branch of ZAP. Be aware that
I don't provide build.zig.zon
snippets or tagged releases for it for
the time being. If you know what you are doing, that shouldn't stop you
from using it with zig master though.zig build run-docserver
to serve them locally.-Dopenssl
flag or the environment variable ZAP_USE_OPENSSL=true
:.openssl = true,
(in dependent projects' build.zig,
b.dependency("zap" .{...})
)ZAP_USE_OPENSSL=true zig build https
zig build -Dopenssl=true https
I recommend checking out Endpoint-based examples for more realistic use cases. Most of the examples are super stripped down to only include what's necessary to show a feature.
NOTE: To see API docs, run zig build run-docserver
. To specify a custom
port and docs dir: zig build docserver && zig-out/bin/docserver --port=8989 --docs=path/to/docs
.
build.zig
now uses the new Zig package
manager for its C-dependencies, no git submodules anymore./users
endpoint for performing PUT/DELETE/GET/POST operations and listing
users, together with a simple frontend to play with. It also introduces a
/stop
endpoint that shuts down Zap, so memory leak detection can be
performed in main().GeneralPurposeAllocator
to report memory leaks when ZAP is shut down.
The StopEndpoint just stops ZAP when
receiving a request on the /stop
route.zap.Middleware.EndpointHandler
. Mixing endpoints
in your middleware chain allows for usage of Zap's authenticated endpoints and
your custom endpoints. Since Endpoints use a simpler API, you have to use
r.setUserContext()
and r.getUserContext()
with the request if you want to
access the middleware context from a wrapped endpoint. Since this mechanism
uses an *anyopaque
pointer underneath (to not break the Endpoint API), it is
less type-safe than zap.Middleware
's use of contexts.setUserContext()
and getUserContext()
, you can, of course use those two in
projects that don't use zap.Endpoint
or zap.Middleware
, too, if you
really, really, absolutely don't find another way to solve your context
problem. We recommend using a zap.Endpoint
inside of a struct that
can provide all the context you need instead. You get access to your
struct in the callbacks via the @fieldParentPtr()
trick that is used
extensively in Zap's examples, like the endpoint
example.r.sendError(err, status_code)
when you catch an error and a stack trace
will be returned to the client / browser.-Dopenssl=true
or the environment
variable ZAP_USE_OPENSSL
set to true
and requires openssl dev dependencies
(headers, lib) to be installed on the system.ZAP_USE_OPENSSL=true zig build run-https
OR like this: zig build -Dopenssl=true run-https
zap.Router
to dispatch to handlers by HTTP path.I'll continue wrapping more of facil.io's functionality and adding stuff to zap to a point where I can use it as the JSON REST API backend for real research projects, serving thousands of concurrent clients.
Claiming to be blazingly fast is the new black. At least, Zap doesn't slow you down and if your server performs poorly, it's probably not exactly Zap's fault. Zap relies on the facil.io framework and so it can't really claim any performance fame for itself. In this initial implementation of Zap, I didn't care about optimizations at all.
But, how fast is it? Being blazingly fast is relative. When compared with a simple GO HTTP server, a simple Zig Zap HTTP server performed really good on my machine (x86_64-linux):
Update: Thanks to @felipetrz, I got to test against more realistic Python
and Rust examples. Both python sanic
and rust axum
were easy enough to
integrate.
Update: I have automated the benchmarks. See blazingly-fast.md for more information. Also, thanks to @alexpyattaev, the benchmarks are fairer now, pinning server and client to specific CPU cores.
Update: I have consolidated the benchmarks to one good representative per language. See more details in blazingly-fast.md. It contains rust implementations that come pretty close to Zap's performance in the simplistic testing scenario.
So, being somewhere in the ballpark of basic GO performance, zig zap seems to be ... of reasonable performance 😎.
I can rest my case that developing ZAP was a good idea because it's faster than both alternatives: a) staying with Python, and b) creating a GO + Zig hybrid.
See more details in blazingly-fast.md.
ZAP is very robust. In fact, it is so robust that I was confidently able to only work with in-memory data (RAM) in all my ZAP projects so far: over 5 large online research experiments. No database, no file persistence, until I hit "save" at the end 😊.
So I was able to postpone my cunning data persistence strategy that's similar to a mark-and-sweep garbage collector and would only persist "dirty" data when traffic is low, in favor of getting stuff online more quickly. But even if implemented, such a persistence strategy is risky because when traffic is not low, it means the system is under (heavy) load. Would you confidently NOT save data when load is high and the data changes most frequently -> the potential data loss is maximized?
To answer that question, I just skipped it. I skipped saving any data until receiving a "save" signal via API. And it worked. ZAP just kept on zapping. When traffic calmed down or all experiment participants had finished, I hit "save" and went on analyzing the data.
Handling all errors does pay off after all. No hidden control flow, no hidden errors or exceptions is one of Zig's strengths.
To be honest: There are still pitfalls. E.g. if you request large stack sizes for worker threads, Zig won't like that and panic. So make sure you don't have local variables that require tens of megabytes of stack space.
See the StopEndpoint in the
endpoint example. That example uses ZIG's awesome
GeneralPurposeAllocator
to report memory leaks when ZAP is shut down. The
StopEndpoint
just stops ZAP when receiving a request on the /stop
route.
You can use the same strategy in your debug builds and tests to check if your code leaks memory.
Make sure you have zig 0.13.0 installed. Fetch it from here.
$ git clone https://github.com/zigzap/zap.git
$ cd zap
$ zig build run-hello
$ # open http://localhost:3000 in your browser
... and open http://localhost:3000 in your browser.
Make sure you have the latest zig release (0.13.0) installed. Fetch it from here.
If you don't have an existing zig project, create one like this:
$ mkdir zaptest && cd zaptest
$ zig init
$ git init ## (optional)
Note: Nix/NixOS users are lucky; you can use the existing flake.nix
and run
nix develop
to get a development shell providing zig and all
dependencies to build and run the GO, python, and rust examples for the
wrk
performance tests. For the mere building of zap projects,
nix develop .#build
will only fetch zig 0.11.0. TODO: upgrade to latest zig.
With an existing Zig project, adding Zap to it is easy:
build.zig.zon
build.zig
In your zig project folder (where build.zig
is located), run:
zig fetch --save "git+https://github.com/zigzap/zap#v0.9.1"
Then, in your build.zig
's build
function, add the following before
b.installArtifact(exe)
:
const zap = b.dependency("zap", .{
.target = target,
.optimize = optimize,
.openssl = false, // set to true to enable TLS support
});
exe.root_module.addImport("zap", zap.module("zap"));
From then on, you can use the Zap package in your project. Check out the examples to see how to use Zap.
You can change the URL to Zap in your build.zig.zon
0.0.9
zap
Go to the release page. Every release
will state its version number and also provide instructions for changing
build.zig.zon
and build.zig
.
See here.
At the current time, I can only add to zap what I need for my personal and professional projects. While this happens blazingly fast, some if not all nice-to-have additions will have to wait. You are very welcome to help make the world a blazingly fast place by providing patches or pull requests, add documentation or examples, or interesting issues and bug reports - you'll know what to do when you receive your calling 👼.
Check out CONTRIBUTING.md for more details.
See also introducing.md for more on the state and progress of this project.
We now have our own ZAP discord server!!!
You can also reach me on the zig showtime discord server under the handle renerocksai (renerocksai#1894).
Being blazingly fast requires a constant feed of caffeine. I usually manage to provide that to myself for myself. However, to support keeping the juices flowing and putting a smile on my face and that warm and cozy feeling into my heart, you can always buy me a coffee ☕. All donations are welcomed 🙏 blazingly fast! That being said, just saying "hi" also works wonders with the smiles, warmth, and coziness 😊.
You build and run the examples via:
$ zig build [EXAMPLE]
$ ./zig-out/bin/[EXAMPLE]
... where [EXAMPLE]
is one of hello
, routes
, serve
, ... see the list of
examples above.
Example: building and running the hello example:
$ zig build hello
$ ./zig-out/bin/hello
To just run an example, like routes
, without generating an executable, run:
$ zig build run-[EXAMPLE]
Example: building and running the routes example:
$ zig build run-routes
const std = @import("std");
const zap = @import("zap");
fn on_request(r: zap.Request) void {
if (r.path) |the_path| {
std.debug.print("PATH: {s}\n", .{the_path});
}
if (r.query) |the_query| {
std.debug.print("QUERY: {s}\n", .{the_query});
}
r.sendBody("<html><body><h1>Hello from ZAP!!!</h1></body></html>") catch return;
}
pub fn main() !void {
var listener = zap.HttpListener.init(.{
.port = 3000,
.on_request = on_request,
.log = true,
});
try listener.listen();
std.debug.print("Listening on 0.0.0.0:3000\n", .{});
// start worker threads
zap.start(.{
.threads = 2,
.workers = 2,
});
}