Enhance log parsing and troubleshoot analysis: robust URL handling, tail-by-minutes, improved select/release matching and reporting

[mouxinqq]

Motivation

• Make URL matching and worker display robust to https, hostnames and optional ports and consistently strip schemes for human-friendly reporting.
• Support tailing by a recent time window (minutes) using log timestamps instead of approximated line counts.
• Improve select/release correlation using available IDs and a safe fallback by worker+time to reduce false unmatched/miscounts.
• Fix brittle defaults for missing dates/times and extend logging level recognition to include DEBUG.

Description

• Introduced a reusable URL_RE pattern and replaced many literal http://... regexes across log_parser.py, stat_cache_hitrate.py, and analyzer modules to accept optional http(s)://, hostnames, IPs and optional ports.
• Added _strip_scheme utility in multiple modules and used it to present worker URLs without http(s):// in reports and diagnostics.
• Replaced hardcoded fallback years/dates in complete_time_arg with datetime.now().year and dynamic defaults.
• Added filter_file_by_recent_minutes and changed CLI --tail handling in troubleshoot.py to accept Nm (minutes) or line counts; Nm now filters by timestamp window.
• Reworked select/release matching in log_parser.match_select_release: build a key via request_id (preferred) or req_id/trace_id/session_id, report id_coverage, collect untracked_selects, and add a worker_time_fallback match within a configurable fallback_window_s. Added _parse_ts_safe and safer timestamp handling.
• Extended match outputs (untracked_selects, id_coverage, match_method) and updated analyzers/reports to respect cases with no correlatable IDs (show Delta as N/A and avoid misreporting).
• Improved token/counter analysis: included additional H11 lines in match_select_release usage, expanded token diagnostics in load analyzer, and added token-leak diagnoses.
• Broadened LEVEL_RE to include DEBUG and tightened several regexes to avoid accidental captures; updated parsing and formatting usages across trace.py, health.py, cache.py, load.py, and stat_cache_hitrate.py.
• Miscellaneous: added imports (re, timedelta) where needed and small refactors to keep parsing robust and human-readable.

Testing

• No automated tests were added or executed as part of this change.

---

Codex Task