Julia Compiler Performance Monitoring & Benchmarking

This repository contains the code and runtime assets for Julia compiler performance monitoring.

Key pieces:

• site: the Rust web UI and HTTP API.
• src: the Julia orchestrator that ingests upstream data and keeps the site process running.
• run.jl: the production entrypoint, which runs the orchestrator in install mode.

Container Image

The repository root Dockerfile builds the full production image. It compiles the embedded prod_site Rust binary, installs the Julia project, and starts the app through run.jl --install.

The production Terraform path runs that image as an unprivileged service user with no-new-privileges and no ambient Linux capabilities. The runtime data directory on the host is owned by that fixed UID/GID automatically during instance bootstrap.

The container expects persistent runtime data under one directory:

• julia.db: the SQLite database.
• .state/: the checkpoint files used by the Julia orchestrator.

Build the image locally with:

docker build -t rustc-perf:latest .

Run it locally against an existing database directory with:

docker run --rm -p 2346:2346 \
	-v "$PWD/runtime-data:/var/lib/rustc-perf" \
	rustc-perf:latest

The first start requires runtime-data/julia.db and both checkpoint files (runtime-data/.state/julia_last_processed_commit.txt and runtime-data/.state/reports_last_processed_commit.txt) to already exist — in production all of them are restored from the S3 backup. The orchestrator fails loudly if any of them is missing rather than guessing a starting point.

Public Database Queries

Production exposes the live SQLite database through a read-only Datasette UI and API at /db/. It supports schema browsing, SELECT queries, and JSON/CSV exports; it does not expose a SQLite network port or database-file downloads. Public queries are limited to one second and 1,000 rows, and may briefly observe either side of an ingestion transaction.

To publish and deploy a new version to production, run infra/terraform/deploy.sh: it builds the image for linux/amd64, pushes it to ECR, takes a pre-deploy backup, and replaces the instance with the new digest-pinned image.

Additional documentation on running and setting up the frontend and backend can be found in the README files in the collector and site directories.

For the step-by-step production bring-up checklist, including deploys, backups, and restore operations, see docs/production-deployment.md.

License

The code of this repository is licensed under the MIT license, managed by the Reuse Specification.