v6.0.2 - Parallel-Scan Stability

RedAmon v6.0.2 - Parallel-Scan Stability

A stability release: fixes the orchestrator freeze and the client-side crash that could take down the UI while many scans run in parallel, plus over-conservative memory admission that falsely rejected new scans.

🧊 Orchestrator no longer freezes under load

• Synchronous Docker calls on the single event loop (status polls, scan spawns, a per-log-line container.reload()) moved onto dedicated thread pools, with the liveness check throttled
• Health checks, status polls, and new scan starts stay responsive even during a GraphQL scan plus many parallel partial recons
• Covers recon, partial recon, GVM, GitHub-hunt, and TruffleHog

💥 Client-side crash fixed

• A rejected scan start returned the memory governor's structured payload as an object, which the UI rendered as a React child ("Objects are not valid as a React child") and took the whole page down
• All scan start routes now normalize the error to a string; a new root error boundary catches any other render error instead of white-screening

🧮 Accurate memory admission

• Scans reserved a flat 4 GB each (~26x the measured ~150 MB peak), so a few parallel partials exhausted the reserved budget while 10+ GB was still free
• Envelopes right-sized: partial recon 512 MB, full recon 1.5 GB
• Webapp requests to a stalled orchestrator now time out cleanly instead of hanging

⚙️ Defaults

• Concurrent-scan ceilings raised to 30 (per-user and global)

Full changelog: v5.1.0...v6.0.2