Filter usage timing overview to benchmark records

[yoshifuminakamura]

Summary

• limit the execution timing overview to known benchmark app records with a concrete system
• exclude diagnostic or malformed Result JSON records that do not carry timing, profiling, or build-cache telemetry
• clarify the overview as an operator view for trigger scope/frequency and CI/build-cache flow decisions
• split observed run time into regular and profiled runs when Result JSON timing is available
• record estimate-job elapsed time in Estimate JSON and show estimate runtime when available

Validation

• .venv/bin/python -m pytest result_server/tests/test_performance_telemetry.py result_server/tests/test_usage_report_view.py result_server/tests/test_usage_route.py result_server/tests/test_portal_list_templates.py
• bash scripts/tests/test_estimation_run_timing.sh
• shellcheck -S error scripts/estimation/run.sh scripts/tests/test_estimation_run_timing.sh
• .venv/bin/python result_server/tests/run_result_server_tests.py
• python3 scripts/tests/check_text_integrity.py
• python3 scripts/tests/check_commit_messages.py
• python3 scripts/tests/check_app_contracts.py
• git diff --check