Skip to main content
All projects
infrastructure

Raft Task Queue

Distributed task queue built in Rust, with Raft consensus implemented from the original paper: leader election, log replication, an Axum REST API, and a real-time web dashboard.

Sole engineer

tasks lost on node failure
0
tasks/sec, single node
5M
task latency
200 ns
Rust crates
5

Throughput and latency measure the in-memory single-node queue with Criterion.rs: 100 iterations, 95% confidence intervals. These are not replicated-cluster throughput measurements.

Loading...

The challenge

A task queue that survives node failure can't just lock a database — it needs real distributed consensus so work is never lost or run twice when a node dies.

What I built

Written from the original Raft paper rather than a tutorial or an off-the-shelf crate: leader election and log replication across 5 Rust crates keep the cluster serving tasks through node failures with zero task loss. The queue itself benchmarks at 5M tasks/sec with 200 ns latency under Criterion.rs (100 iterations, 95% confidence intervals). That number is the in-memory single-node path — the queue's own cost, not a replicated-cluster throughput claim.

Have something in mind?

I'm open to backend and platform roles, contract or permanent. Based in the EU, available on-site in Belgium and the Netherlands or remotely worldwide.

Download CV