docs: add rate limiting guide

[CharlieTLe]

Summary

Adds a new guide at docs/guides/rate-limiting.md documenting the write-path rate limiters, which were previously only partially documented in docs/configuration/arguments.md and the auto-generated config reference.

The new page covers:

• Per-tenant ingestion limiter — token-bucket semantics, what totalN counts (samples + exemplars + metadata), and the 10s recheck for runtime overrides.
• local vs global strategy — including the load-balancing assumption and the fact that ingestion_burst_size is not divided by N under the global strategy.
• Sizing guidance — burst must be ≥ the largest expected single push.
• Native histogram limiter — separate token bucket, applied after the main limiter, so a histogram-only rejection still ingests floats / exemplars / metadata.
• Per-instance limits — -distributor.instance-limits.max-ingestion-rate and -ingester.instance-limits.max-ingestion-rate (EWMA-based, HTTP 503 / gRPC error).
• Errors and observability — the exact response codes and messages, retry_on_http_429 behavior, and the relevant cortex_distributor_*, cortex_discarded_*, and cortex_overrides metrics.

Also adds a cross-link to the new guide from the per-tenant limits section in docs/configuration/arguments.md.

Docs-only change — no CHANGELOG entry.

Test plan

• Render the docs locally and verify the new page appears under Guides and that the cross-link from arguments.md resolves.
• Spot-check internal links: ./overrides.md, ./overrides-exporter.md, ../configuration/arguments.md.

[SungJin1212]

thanks!