docs: improve performance tuning and profiling guidance

[evenyag]

What's Changed in this PR

• Expands the performance tuning tips page with EXPLAIN ANALYZE VERBOSE guidance for debugging slow queries, including common scan, filter, cache, and metadata findings.
• Updates cache tuning guidance with current cache metrics, cache type labels, index cache allocation, range result and prefilter result caches, and index cache sizing tips.
• Adds ingestion troubleshooting guidance for throughput, request latency, datanode write pressure, flush performance, write distribution, and write buffer sizing.
• Adds CPU and memory profiling collection guidance for nodes with high resource usage, including platform availability, node-local collection, flamegraph overhead, and issue reporting details.
• Refreshes the Profiling Tools HTTP endpoint reference with current CPU and memory profiling endpoints, output formats, heap profiling status and activation, gdump, and symbolication.
• Updates the region engine configuration reference with corrected cache default descriptions and the range_result_cache_size and prefilter_result_cache_size options.

Checklist

• Please confirm that all corresponding versions of the documents have been revised.
• Please ensure that the content in sidebars.ts matches the current document structure when you changed the document structure.
• This change requires follow-up update in localized docs.

[fengjiachun]

En part LGTM