refactor(config): config consolidation for Configra

[KooshaPari]

Stage 1 config consolidation: extracted hardcoded config to configra-config crate

Summary

Replaces hardcoded configuration defaults (port 8080, log level "info", idempotency TTL 86400s) with canonical defaults sourced from the new configra-config meta-configuration crate.

Changes

New: crates/configra-config/

• ConfigraConfig struct with three sub-config groups: ServiceConfig, IdempotencyConfig, WatcherConfig
• Loading from: environment variables (CONFIGRA_* prefix), config files (TOML/JSON), or programmatic builder
• All fields have documented defaults mirroring the previously hardcoded values
• Example config files: configra.example.toml, configra.example.json

Modified: crates/pheno-config/src/lib.rs

• load_from_env() now uses ConfigraConfig::default().service.default_port / .default_log_level instead of hardcoded 8080 / "info"
• ConfigBuilder::new() same treatment

Modified: crates/settly/src/adapters/idempotency.rs

• InMemoryIdempotencyStore::default_ttl() uses ConfigraConfig::default().idempotency.default_ttl_secs instead of hardcoded 86_400

Documentation

• docs/CONFIG.md — full reference of every config key, its type, default, env var, and origin
• Crate-level docs updated where defaults were previously documented as literals

Other

• Workspace member list updated to include configra-config
• Example files (examples/cascade.rs, quickstart.rs, validation.rs) modernized for the current API

Verification

cargo check --workspace passes. All defaults remain semantically identical — no behavioral changes.

Closes config consolidation phase (ADR-031 follow-up).

[gemini-code-assist]

Code Review

This pull request introduces the configra-config crate to centralize meta-configuration defaults (such as ports, log levels, and idempotency TTLs) across the workspace, updating pheno-config and settly to utilize these shared defaults. The review feedback highlights opportunities to improve robustness and performance: specifically, making ConfigraConfig::load fallible to prevent silently ignoring malformed configuration files, filtering out empty environment variables in from_env to ensure proper default fallbacks, and optimizing ConfigBuilder::new to avoid redundant default configuration instantiations and string clones.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[gemini-code-assist]

The load function silently falls back to ConfigraConfig::default() if from_file fails (e.g., due to a syntax error in the TOML/JSON file). This can lead to extremely hard-to-debug issues in production where a malformed configuration file is silently ignored, and the application runs with default settings instead of the intended overrides. Consider providing a fallible try_load function that propagates the ConfigError, allowing the application to fail-fast on startup.

Suggested change

	pub fn load(path: Option<&Path>) -> Self {
	match path {
	Some(p) if p.exists() => Self::from_file(p).unwrap_or_default(),
	_ => Self::from_env(),
	}
	}
	pub fn load(path: Option<&Path>) -> Result<Self> {
	match path {
	Some(p) if p.exists() => Self::from_file(p),
	_ => Ok(Self::from_env()),
	}
	}

[gemini-code-assist]

If the environment variable CONFIGRA_DEFAULT_LOG_LEVEL or CONFIGRA_DB_PATH_TEMPLATE is set to an empty string, std::env::var will return Ok(""), and from_env() will use the empty string instead of falling back to the default value. In contrast, pheno-config explicitly filters out empty strings (using if !v.is_empty()). To ensure consistency and prevent empty configuration values, consider filtering out empty strings using .filter(|v| !v.is_empty()) before unwrapping.

        let default_log_level = std::env::var(env_name("DEFAULT_LOG_LEVEL"))
            .ok()
            .filter(|v| !v.is_empty())
            .unwrap_or_else(|| "info".to_owned());

        let db_path_template = std::env::var(env_name("DB_PATH_TEMPLATE"))
            .ok()
            .filter(|v| !v.is_empty())
            .unwrap_or_else(|| "/var/lib/{name}.db".to_owned());

[gemini-code-assist]

In ConfigBuilder::new(), ConfigraConfig::default() is called twice, which unnecessarily constructs the entire configuration structure twice and performs redundant heap allocations for the default strings. Additionally, default_log_level is cloned. By instantiating ConfigraConfig::default() once, you can avoid the duplicate construction and directly move the default_log_level string without cloning it.

Suggested change

	pub fn new() -> Self {
	Self {
	url: None,
	port: 8080,
	log_level: String::from("info"),
	port: ConfigraConfig::default().service.default_port,
	log_level: ConfigraConfig::default().service.default_log_level.clone(),
	db_path: None,
	feature_flags: Vec::new(),
	}
	pub fn new() -> Self {
	let defaults = ConfigraConfig::default();
	Self {
	url: None,
	port: defaults.service.default_port,
	log_level: defaults.service.default_log_level,
	db_path: None,
	feature_flags: Vec::new(),
	}
	}

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3c9780b263

ℹ️ About Codex in GitHub

Codex has been enabled to automatically review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

When you sign up for Codex through ChatGPT, Codex can also answer questions or update the PR, like "@codex address that feedback".

[chatgpt-codex-connector]

Use a toolchain that satisfies pheno-config's MSRV

In the new .github/workflows/ci.yml MSRV job, CI installs Rust 1.75 and then runs cargo check --workspace; I inspected the workspace manifests and crates/pheno-config/Cargo.toml declares rust-version = "1.82", so this job will fail for every PR/push that checks the full workspace before code is built. Either run this workspace check on 1.82 or exclude/split the crate that requires the newer compiler.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Add a registry version for configra-config

This new dependency is path-only in a publishable crate, while the added release workflow publishes pheno-config; I verified the packaging semantics with cargo package -p pheno-config --allow-dirty --no-verify, which aborts with all dependencies must have a version requirement specified when packaging because Cargo removes path specs for packaged dependencies. The same path-only dependency was added to settly, so tag releases cannot publish these crates unless configra-config gets a version requirement and is published before its dependents.

Useful? React with 👍 / 👎.

[chatgpt-codex-connector]

Default missing top-level config sections

Partial TOML/JSON files are documented and tested as valid, but #[serde(default)] is only on the nested section structs; missing root fields like idempotency and watcher are still required during ConfigraConfig::from_file deserialization. A file containing only [service] default_port = 5000 therefore errors with a missing top-level section instead of filling defaults, so add defaults on the root struct/fields as well.

Useful? React with 👍 / 👎.