fix(workflow): prevent checkpoint expiry for paused workflows (#3231)

[mrveiss]

Summary

Two independent checkpoint systems both used TTLs too short for human-in-the-loop pauses:

• WorkflowCheckpointManager (error_handler.py): TTL extended 7 days → 30 days. New refresh_ttl() method resets the countdown on every resume, so a workflow paused near its deadline gets a fresh 30-day window when the operator acts.
• AsyncRedisSaver (graph.py + ssot_config.py): Default TTL extended 24 hours → 30 days (1440 → 43200 minutes). In-code fallback updated to match.
• workflow_executor.py: _apply_checkpoints() now calls checkpoint_manager.refresh_ttl() immediately on resume.
• Tests: 4 new tests covering the 30-day constant, TTL being set on save, TTL refresh on resume, and Redis errors in refresh_ttl being logged (never raised).

Closes #3231

Test plan

• pytest autobot-backend/orchestration/error_handler_test.py -v — all pass
• Workflow paused for longer than previous 7-day / 24-hour TTL can still be resumed
• AUTOBOT_REDIS_CHECKPOINT_TTL_MINUTES env var overrides the default

[github-actions]

✅ SSOT Configuration Compliance: Passing

🎉 No hardcoded values detected that have SSOT config equivalents!

[github-actions]

AutoBot Phase Validation Results

System Maturity: 0.0%

Phase Status:

Recommendations:

[mrveiss]

Code review

No issues found. Checked for bugs and CLAUDE.md compliance.

🤖 Generated with Claude Code

_{- If this code review was useful, please react with 👍. Otherwise, react with 👎.}