Problem

All retry loops use fixed intervals. For transient failures (network blips, Docker API errors), fixed intervals waste time. For persistent failures, they don't fail fast enough.

Files: src/container.py (wait_until_healthy), src/initialize.py (get_mylogin_cnf)

Fix

for attempt in range(retries):
    delay = min(base_interval * (2 ** attempt), max_interval)
    time.sleep(delay)

Acceptance Criteria

• wait_until_healthy() retry loop uses exponential backoff (base_interval * 2^attempt, capped at max_interval)
• get_mylogin_cnf() retry loop in initialize.py uses exponential backoff
• Backoff parameters (base, max) are configurable class-level constants