Summary

Log timestamps appear in mixed timezone formats (UTC-like and local timezone) within the same operational workflow, causing confusion during incident triage.

Environment

• OpenClaw 2026.3.2
• Windows 11
• User timezone: Europe/Moscow (UTC+3)

Observed behavior

Operators see entries like 08:31 and 11:31 referring to the same runtime period, depending on log path/formatter.

This is especially confusing during outages where minute-level ordering matters.

Expected

• Consistent timezone policy across all emitted logs (pretty + plain + structured).
• If mixed modes are intentional, logs should clearly include timezone offset for every line.

Actual

• Some outputs appear UTC-like while others are local-time, without consistent explicit offset in all paths.

Impact

• wrong incident timeline reconstruction
• extra MTTR during crash-loop debugging

Suggested direction

• unify formatter timezone source,
• or force explicit ISO timestamp with offset in all log modes,
• and document the behavior per command mode (logs --plain, --local-time, live gateway console, etc.).