Two changes combined:
1. historical-query-slow-connection: Migrate all historical query pages
to read_sql_df_slow with semaphore concurrency control (max 3),
raise DB slow timeout to 300s, gunicorn timeout to 360s, and
unify frontend timeouts to 360s for all historical pages.
2. hold-resource-history-dataset-cache: Convert hold-history and
resource-history from multi-query to single-query + dataset cache
pattern (L1 ProcessLevelCache + L2 Redis parquet/base64, TTL=900s).
Replace old GET endpoints with POST /query + GET /view two-phase
API. Frontend auto-retries on 410 cache_expired.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Use Daily Trend as single source of truth for On Hold and New Hold KPI cards
instead of separate snapshot SQL queries, eliminating value mismatches. Fix
timezone bug in default date range (toISOString UTC offset), add 1st-of-month
fallback to previous month, replace Hold Type radio buttons with select dropdown,
reorder/relabel summary cards with 累計 prefix, add job-query MultiSelect for
equipment filter, and fix heatmap chart X-axis overlap with visualMap legend.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
New independent report page based on DWH.DW_MES_HOLDRELEASEHISTORY providing
historical hold/release performance analysis. Includes daily trend with Redis
caching, reason Pareto with click-to-filter, duration distribution with
click-to-filter, multi-select record type filter (new/on_hold/released),
workcenter-group mapping via memory cache, and server-side paginated detail
table. All 32 backend tests passing.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
