circus/docs/COMPARISON.md

FC vs Hydra

FC is not exactly a competitor, but a direct replacement to Hydra with a limited set of features that are prioritized. While FC hopes to be a complete replacement, not everything can be considered in-scope for a team of FOSS developers without funding. Should you create any issues, your feature may or may not be implemented depending on our own needs. Please keep that in mind.

Below document contains a "feature matrix" of concepts and features that we've decided to think about. Not all of them will be fully implemented (i.e., we are a-okay with being less powerful in some regards) but FC does aim to be better than Hydra in the long term, through different means; reliability, UX and performance.

Executive Summary

FC currently implements more or less 50% of Hydra's core features, and has various improvements over Hydra's architecture. As of writing, some gaps (such as the plugin architecture, VCS diversity and notification integrations) remain.

Strengths

• Modern Rust codebase with better error handling
• Simpler, more maintainable architecture (5 Rust crates vs Perl & C++ monolith)
• Better API-first design with proper REST structure
• User management with argon2 password hashing and granular RBAC
• Cleaner database schema (16 tables vs 70+)
• Better Nix Flake support from day one
• Improved & tested declarative jobsets

Weaknesses

• Limited VCS support (Git only in FC vs 6 types in Hydra)
• No plugin architecture for extensibility
• Missing several notification integrations (Slack, BitBucket, etc.)
• No declarative project specification (coming soon)
• No coverage/build metrics collection

TODO: add a better comparison matrix