feat(health): Configurable tunnel health monitoring and reconnect thresholds (#44)

[Vonarian]

Resolves #44

Summary of Changes

• Configurable Health Monitoring Flags:

  • --health-interval <n> (AETHER_HEALTH_INTERVAL, default 20s): Controls continuous background check frequency. Set to 0 to disable.
  • --health-fails <n> (AETHER_HEALTH_MAX_FAILS, default 2): Maximum consecutive failed probes before declaring tunnel dead and initiating reconnect.
  • --health-timeout <n> (AETHER_HEALTH_TIMEOUT, default 5s): Timeout per health probe check.
  • --health-url <url> (AETHER_HEALTH_PROBE_URL, default http://www.gstatic.com/generate_204): Custom health probe URL.
  • --reconnect-secs <n> (AETHER_MASQUE_RECONNECT_SECS / AETHER_WG_RECONNECT_SECS, default 2s): Reconnect delay after tunnel exit or failure.

• Observatory-Style Health Monitor:

  • Implemented background health monitor task in tunnelping.rs that probes active netstack instances periodically.
  • Measures and logs RTT latency per check (e.g. [+] tunnel health check passed (rtt 142ms)).
  • Automatically triggers clean tunnel reconnect upon consecutive_fails >= max_fails.

• Integration:

  • Integrated into run_masque_tunnel, run_wireguard_tunnel, and run_warp_in_warp via tokio::select!.

[Copilot]

Pull request overview

This PR adds configurable, continuous tunnel health monitoring (with timeouts and failure thresholds) and integrates it into MASQUE, WireGuard, and warp-in-warp runtimes, while also refining logging verbosity controls.

Changes:

• Added HealthConfig + background health monitor task that periodically probes a configurable HTTP URL and triggers failure on consecutive probe errors.
• Wired the health monitor into tunnel run loops via tokio::select! and exposed config via CLI/env (--health-*, AETHER_HEALTH_*), plus unified --reconnect-secs for MASQUE and WireGuard.
• Adjusted several hot-path logs from debug to trace and introduced AETHER_LOG/--log-level + multi-level -v verbosity parsing.

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
aether/src/wireguard.rs	Reduces noise by downgrading encapsulation/decapsulation errors to trace.
aether/src/tunnelping.rs	Introduces probe target parsing, health config/env wiring, and the background health monitor + tests.
aether/src/quic.rs	Downgrades verbose per-packet logs to trace to reduce default log volume.
aether/src/noize.rs	Downgrades obfuscation send logs to trace to reduce default log volume.
aether/src/masque_h2.rs	Downgrades capsule send/parse logs to trace.
aether/src/main.rs	Integrates the health monitor into MASQUE/WireGuard/warp-in-warp lifecycles and switches logging init to AETHER_LOG.
aether/src/cli.rs	Adds health-monitoring CLI flags, new logging flags, and unit tests for env/arg parsing.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Vonarian]

Addressed PR review feedback in commit 97892b1:

1. Task Cancellation Safety: Borrowed join handles (&mut tunnel_task, &mut socks_task, ref mut health_task) in tokio::select! and added explicit .abort() calls on all tasks before returning in run_masque_tunnel, run_wireguard_tunnel, and run_warp_in_warp.
2. Probe URL Scheme Validation: Updated parse_probe_url to return Result<ProbeTarget> and explicitly reject https:// schemes with an error (https:// health probe URLs are not supported; please use an http:// URL) to avoid implicit scheme downgrades.
3. Test Env Hermeticity: Added std::env::remove_var calls at the end of test_cli_parse_log_levels and test_cli_parse_health_options.