Add Redis-backed Rate Limiter (.NET 8+)

Summary

Implement System.Threading.RateLimiting.RateLimiter backed by Redis for distributed rate limiting across multiple app instances.

Problem

.NET 8's RateLimiting middleware is in-memory by default. In a multi-instance deployment, each instance has its own counter — so a limit of 100 req/min becomes 100×N. Redis-backed rate limiting solves this.

Proposed Algorithms

Fixed Window

services.AddRedisFixedWindowRateLimiter("api", options =>
{
    options.Window = TimeSpan.FromMinutes(1);
    options.PermitLimit = 100;
});

Sliding Window

services.AddRedisSlidingWindowRateLimiter("api", options =>
{
    options.Window = TimeSpan.FromMinutes(1);
    options.PermitLimit = 100;
    options.SegmentsPerWindow = 6;
});

Implementation

• Use `StringIncrementAsync` with TTL for fixed window
• Use Lua script for sliding window (atomic check-and-increment)
• Consider a separate package: `StackExchange.Redis.Extensions.RateLimiting`

Considerations

• SE.Redis 2.12 has `StringGcraRateLimitAsync` — a built-in GCRA algorithm
• Depends on #634 (StringIncrementAsync)

Start with issue #634 and the existing Redis APIs, especially StringIncrementAsync and the SE.Redis 2.12 GCRA option. Compare the proposed fixed-window and sliding-window entry points, including their package boundary and atomicity requirements. Done means distributed rate limits work across multiple app instances for both proposed algorithms.