Redis (StackExchange.Redis)

Overview

Redis is an in-memory data structure store used as a distributed cache, message broker, and database. StackExchange.Redis is the primary .NET client library, providing a high-performance, multiplexed connection to Redis with support for all Redis data types: strings, hashes, lists, sets, sorted sets, streams, and pub/sub channels.

The library is built around the ConnectionMultiplexer, which manages connections efficiently and is designed to be shared as a singleton across the application. It supports both synchronous and asynchronous operations, pipelining, Lua scripting, transactions, and cluster mode.

Install via NuGet: dotnet add package StackExchange.Redis

Connection Setup and DI Registration

using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Hosting;
using StackExchange.Redis;

var builder = Host.CreateApplicationBuilder(args);

// Register ConnectionMultiplexer as a singleton
builder.Services.AddSingleton<IConnectionMultiplexer>(sp =>
{
    var configuration = ConfigurationOptions.Parse(
        builder.Configuration.GetConnectionString("Redis")!);
    configuration.AbortOnConnectFail = false;
    configuration.ConnectRetry = 3;
    configuration.ConnectTimeout = 5000;
    configuration.SyncTimeout = 5000;
    configuration.AsyncTimeout = 5000;
    return ConnectionMultiplexer.Connect(configuration);
});

// Convenience registration for IDatabase
builder.Services.AddScoped<IDatabase>(sp =>
{
    var multiplexer = sp.GetRequiredService<IConnectionMultiplexer>();
    return multiplexer.GetDatabase();
});

var app = builder.Build();
await app.RunAsync();

String Operations (Key-Value)

using StackExchange.Redis;
using System.Text.Json;

public sealed class RedisCacheService
{
    private readonly IDatabase _db;

    public RedisCacheService(IDatabase db)
    {
        _db = db;
    }

    public async Task SetAsync<T>(string key, T value, TimeSpan? expiry = null)
    {
        string json = JsonSerializer.Serialize(value);
        await _db.StringSetAsync(key, json, expiry);
    }

    public async Task<T?> GetAsync<T>(string key)
    {
        RedisValue value = await _db.StringGetAsync(key);
        if (value.IsNullOrEmpty)
        {
            return default;
        }
        return JsonSerializer.Deserialize<T>(value.ToString());
    }

    public async Task<T> GetOrSetAsync<T>(
        string key, Func<Task<T>> factory, TimeSpan expiry)
    {
        RedisValue cached = await _db.StringGetAsync(key);
        if (!cached.IsNullOrEmpty)
        {
            return JsonSerializer.Deserialize<T>(cached.ToString())!;
        }

        T value = await factory();
        string json = JsonSerializer.Serialize(value);
        await _db.StringSetAsync(key, json, expiry);
        return value;
    }

    public async Task<bool> DeleteAsync(string key)
    {
        return await _db.KeyDeleteAsync(key);
    }

    public async Task<long> IncrementAsync(string key, long value = 1)
    {
        return await _db.StringIncrementAsync(key, value);
    }
}

Hash Operations

Hashes store field-value pairs under a single key, ideal for representing objects.

using StackExchange.Redis;

public sealed class UserSessionStore
{
    private readonly IDatabase _db;

    public UserSessionStore(IDatabase db)
    {
        _db = db;
    }

    public async Task SetSessionAsync(string sessionId, UserSession session)
    {
        string key = $"session:{sessionId}";
        HashEntry[] entries = new[]
        {
            new HashEntry("userId", session.UserId),
            new HashEntry("email", session.Email),
            new HashEntry("role", session.Role),
            new HashEntry("loginTime", session.LoginTime.ToString("O")),
            new HashEntry("ipAddress", session.IpAddress)
        };

        await _db.HashSetAsync(key, entries);
        await _db.KeyExpireAsync(key, TimeSpan.FromHours(2));
    }

    public async Task<UserSession?> GetSessionAsync(string sessionId)
    {
        string key = $"session:{sessionId}";
        HashEntry[] entries = await _db.HashGetAllAsync(key);

        if (entries.Length == 0)
        {
            return null;
        }

        var dict = entries.ToDictionary(
            e => e.Name.ToString(),
            e => e.Value.ToString());

        return new UserSession
        {
            UserId = dict["userId"],
            Email = dict["email"],
            Role = dict["role"],
            LoginTime = DateTime.Parse(dict["loginTime"]),
            IpAddress = dict["ipAddress"]
        };
    }

    public async Task UpdateFieldAsync(string sessionId, string field, string value)
    {
        await _db.HashSetAsync($"session:{sessionId}", field, value);
    }
}

public sealed class UserSession
{
    public string UserId { get; set; } = string.Empty;
    public string Email { get; set; } = string.Empty;
    public string Role { get; set; } = string.Empty;
    public DateTime LoginTime { get; set; }
    public string IpAddress { get; set; } = string.Empty;
}

Sorted Sets (Leaderboards)

using StackExchange.Redis;

public sealed class LeaderboardService
{
    private readonly IDatabase _db;
    private const string LeaderboardKey = "game:leaderboard";

    public LeaderboardService(IDatabase db)
    {
        _db = db;
    }

    public async Task AddScoreAsync(string playerId, double score)
    {
        await _db.SortedSetAddAsync(LeaderboardKey, playerId, score);
    }

    public async Task<double> IncrementScoreAsync(string playerId, double increment)
    {
        return await _db.SortedSetIncrementAsync(LeaderboardKey, playerId, increment);
    }

    public async Task<long?> GetRankAsync(string playerId)
    {
        // Rank is zero-based, descending order
        return await _db.SortedSetRankAsync(LeaderboardKey, playerId, Order.Descending);
    }

    public async Task<List<LeaderboardEntry>> GetTopPlayersAsync(int count)
    {
        SortedSetEntry[] entries = await _db.SortedSetRangeByRankWithScoresAsync(
            LeaderboardKey, 0, count - 1, Order.Descending);

        return entries.Select((e, index) => new LeaderboardEntry
        {
            Rank = index + 1,
            PlayerId = e.Element.ToString(),
            Score = e.Score
        }).ToList();
    }
}

public sealed class LeaderboardEntry
{
    public int Rank { get; set; }
    public string PlayerId { get; set; } = string.Empty;
    public double Score { get; set; }
}

Pub/Sub Messaging

using StackExchange.Redis;
using System.Text.Json;

public sealed class RedisEventBus
{
    private readonly IConnectionMultiplexer _multiplexer;

    public RedisEventBus(IConnectionMultiplexer multiplexer)
    {
        _multiplexer = multiplexer;
    }

    public async Task PublishAsync<T>(string channel, T message)
    {
        ISubscriber subscriber = _multiplexer.GetSubscriber();
        string json = JsonSerializer.Serialize(message);
        await subscriber.PublishAsync(RedisChannel.Literal(channel), json);
    }

    public async Task SubscribeAsync<T>(string channel, Action<T> handler)
    {
        ISubscriber subscriber = _multiplexer.GetSubscriber();
        await subscriber.SubscribeAsync(RedisChannel.Literal(channel), (ch, message) =>
        {
            if (!message.IsNullOrEmpty)
            {
                T? value = JsonSerializer.Deserialize<T>(message.ToString());
                if (value is not null)
                {
                    handler(value);
                }
            }
        });
    }

    public async Task UnsubscribeAsync(string channel)
    {
        ISubscriber subscriber = _multiplexer.GetSubscriber();
        await subscriber.UnsubscribeAsync(RedisChannel.Literal(channel));
    }
}

Distributed Locking

using StackExchange.Redis;

public sealed class RedisDistributedLock
{
    private readonly IDatabase _db;

    public RedisDistributedLock(IDatabase db)
    {
        _db = db;
    }

    public async Task<bool> AcquireAsync(string lockKey, string lockValue, TimeSpan expiry)
    {
        return await _db.StringSetAsync(
            $"lock:{lockKey}", lockValue, expiry, When.NotExists);
    }

    public async Task<bool> ReleaseAsync(string lockKey, string lockValue)
    {
        // Only release if we still own the lock (atomic via Lua script)
        const string script = @"
            if redis.call('get', KEYS[1]) == ARGV[1] then
                return redis.call('del', KEYS[1])
            else
                return 0
            end";

        RedisResult result = await _db.ScriptEvaluateAsync(
            script,
            new RedisKey[] { $"lock:{lockKey}" },
            new RedisValue[] { lockValue });

        return (int)result == 1;
    }

    public async Task<T> WithLockAsync<T>(
        string lockKey, TimeSpan timeout, Func<Task<T>> action)
    {
        string lockValue = Guid.NewGuid().ToString();
        var expiry = TimeSpan.FromSeconds(30);
        var start = DateTime.UtcNow;

        while (DateTime.UtcNow - start < timeout)
        {
            if (await AcquireAsync(lockKey, lockValue, expiry))
            {
                try
                {
                    return await action();
                }
                finally
                {
                    await ReleaseAsync(lockKey, lockValue);
                }
            }

            await Task.Delay(50);
        }

        throw new TimeoutException($"Could not acquire lock '{lockKey}' within {timeout}");
    }
}

Rate Limiting with Redis

using StackExchange.Redis;

public sealed class RedisRateLimiter
{
    private readonly IDatabase _db;

    public RedisRateLimiter(IDatabase db)
    {
        _db = db;
    }

    public async Task<RateLimitResult> CheckRateLimitAsync(
        string clientId, int maxRequests, TimeSpan window)
    {
        string key = $"ratelimit:{clientId}";
        long currentCount = await _db.StringIncrementAsync(key);

        if (currentCount == 1)
        {
            await _db.KeyExpireAsync(key, window);
        }

        TimeSpan? ttl = await _db.KeyTimeToLiveAsync(key);

        return new RateLimitResult
        {
            IsAllowed = currentCount <= maxRequests,
            CurrentCount = currentCount,
            Limit = maxRequests,
            RetryAfter = currentCount > maxRequests ? ttl : null
        };
    }
}

public sealed class RateLimitResult
{
    public bool IsAllowed { get; set; }
    public long CurrentCount { get; set; }
    public int Limit { get; set; }
    public TimeSpan? RetryAfter { get; set; }
}

Redis Data Type Selection Guide

Best Practices

1. Register ConnectionMultiplexer as a singleton and reuse it across the entire application; creating multiple multiplexers wastes connections and degrades performance.
2. Set AbortOnConnectFail = false in ConfigurationOptions so the client retries connections gracefully rather than throwing an exception on the first failure.
3. Use KeyExpireAsync on every key that is not meant to live forever to prevent unbounded memory growth in the Redis instance.
4. Use hash operations (HashSetAsync, HashGetAsync) for objects with many fields instead of serializing the entire object as a JSON string, enabling partial field updates.
5. Release distributed locks using a Lua script that checks ownership before deleting to prevent accidentally releasing a lock acquired by another process after expiry.
6. Use FireAndForget command flags on non-critical write operations (e.g., analytics counters) to reduce latency by not waiting for the server acknowledgment.
7. Namespace all keys with a prefix (e.g., "myapp:session:{id}") to avoid collisions when multiple applications share the same Redis instance.
8. Configure SyncTimeout and AsyncTimeout to values appropriate for your latency requirements (typically 1-5 seconds) and handle TimeoutException with retries.
9. Use pipelining by issuing multiple commands before awaiting any results (batch = db.CreateBatch()) to reduce network round trips for bulk operations.
10. Monitor Redis memory usage and eviction policy (maxmemory-policy) in production; use allkeys-lru for cache workloads and noeviction for data that must not be lost.