apalis
Features
- Simple and predictable job handling model.
- Jobs handlers with a macro free API.
- Take full advantage of the
tower
ecosystem of middleware, services, and utilities. - Runtime agnostic - Use tokio, smol etc.
- Optional Web interface to help you manage your jobs.
apalis job processing is powered by tower::Service
which means you have access to the tower
middleware.
apalis has support for:
Source | Crate | Example |
---|---|---|
Cron Jobs | ||
Redis | ||
Sqlite | ||
Postgres | ||
MySQL | ||
Amqp | ||
From Scratch |
Getting Started
To get started, just add to Cargo.toml
[dependencies]
apalis = { version = "0.4", features = ["redis"] }
Usage
use apalis::prelude::*;
use apalis::redis::RedisStorage;
use serde::{Deserialize, Serialize};
use anyhow::Result;
#[derive(Debug, Deserialize, Serialize)]
struct Email {
to: String,
}
impl Job for Email {
const NAME: &'static str = "apalis::Email";
}
/// A function that will be converted into a service.
/// The following signatures are accepted
/// ```rust
/// async fn job(email: Email, ctx: JobContext)
/// async fn job(email: Email, ctx: JobContext) -> anyhow::Result<anyhow::Error>
/// async fn job(email: Email, ctx: JobContext) -> Result<(), my::Error>
/// async fn job(email: Email, ctx: JobContext) -> primitive //eg str
/// async fn job(email: Email, ctx: JobContext) -> impl IntoJobResponse
/// ```
async fn email_service(job: Email, ctx: JobContext) {
info!("Do something");
}
#[tokio::main]
async fn main() -> Result<()> {
std::env::set_var("RUST_LOG", "debug");
env_logger::init();
let redis = std::env::var("REDIS_URL").expect("Missing env variable REDIS_URL");
let storage = RedisStorage::new(redis).await?;
Monitor::new()
.register_with_count(2, move |index| {
WorkerBuilder::new(format!("email-worker-{index}"))
.with_storage(storage.clone())
.build_fn(email_service)
})
.run()
.await
}
Then
//This can be in another part of the program or another application eg a http server
async fn produce_route_jobs(storage: &RedisStorage<Email>) -> Result<()> {
let mut storage = storage.clone();
storage
.push(Email {
to: "[email protected]".to_string(),
})
.await?;
}
Feature flags
- tracing (enabled by default) β Support Tracing
π - redis β Include redis storage
- postgres β Include Postgres storage
- sqlite β Include SQlite storage
- mysql β Include MySql storage
- cron β Include cron job processing
- sentry β Support for Sentry exception and performance monitoring
- prometheus β Support Prometheus metrics
- retry β Support direct retrying jobs
- timeout β Support timeouts on jobs
- limit β πͺ Limit the amount of jobs
- filter β Support filtering jobs based on a predicate
- extensions β Add a global extensions to jobs
Storage Comparison
Since we provide a few storage solutions, here is a table comparing them:
Feature | Redis | Sqlite | Postgres | Sled | Mysql | Mongo | Cron |
---|---|---|---|---|---|---|---|
Scheduled jobs | β | β | β | x | β | x | β |
Retry jobs | β | β | β | x | β | x | β |
Persistence | β | β | β | x | β | x | BYO |
Rerun Dead jobs | β | β | β | x | β | x | x |
How apalis works
sequenceDiagram
participant App
participant Worker
participant Producer
App->>+Producer: Add job to queue
Producer-->>+Worker: Job data
Worker->>+Producer: Update job status to 'running' via Layer
Producer-->>-Worker: Confirmation
Worker->>+App: Notify job started via Layer
loop job execution
Worker-->>-App: Report job progress via Layer
end
Worker->>+Producer: Update job status to 'completed' via Layer
Web UI
If you are running apalis Board, you can easily manage your jobs. See a working rest API example here
Thanks to
tower
- Tower is a library of modular and reusable components for building robust networking clients and servers.- redis-rs - Redis library for rust
- sqlx - The Rust SQL Toolkit
Roadmap
v 1.0
- Refactor the crates structure
- Mocking utilities
- Support for SurrealDB and Mongo
- Lock free for Postgres
- Add more utility layers
- Use extractors in job fn structure
- Polish up documentation
- Improve and standardize apalis Board
- Benchmarks
v 0.4
- Move from actor based to layer based processing
- Graceful Shutdown
- Allow other types of executors apart from Tokio
- Mock/Test Worker
- Improve monitoring
- Add job progress via layer
- Add more sources
v 0.3
- Standardize API (Storage, Worker, Data, Middleware, Context )
- Introduce SQL
- Implement layers for Sentry and Tracing.
- Improve documentation
- Organized modules and features.
- Basic Web API Interface
- Sql Examples
- Sqlx migrations
v 0.2
- Redis Example
- Actix Web Example
Resources
Contributing
Please read CONTRIBUTING.md for details on our code of conduct, and the process for submitting pull requests to us.
Versioning
We use SemVer for versioning. For the versions available, see the tags on this repository.
Authors
- Njuguna Mureithi - Initial work - Njuguna Mureithi
See also the list of contributors who participated in this project.
License
This project is licensed under the MIT License - see the LICENSE.md file for details