Why should you care?

Successful CRM copilot doubles users monthly — architecture must shard hot paths before Postgres connections saturate.

See it live — copy this example

Paste into an ASP.NET Core 8+ / AgentNest project, then run with dotnet run (set your API keys in user-secrets).

// AgentNest.Api/Program.cs — Kestrel limits
builder.WebHost.ConfigureKestrel(o =>
{
    o.Limits.MaxConcurrentConnections = 5000;
    o.Limits.KeepAliveTimeout = TimeSpan.FromMinutes(2);
});
builder.Services.AddStackExchangeRedisCache(o => o.Configuration = redisConn);
builder.Services.AddDbContextPool<AgentNestDbContext>(o => o.UseNpgsql(pgConn), poolSize: 128);

What happened?

• Kestrel connection limits, Redis cache, and DbContext pooling address common ASP.NET bottlenecks as AgentNest tenant count grows.
• Follow the steps below — typing the code yourself is the fastest way to learn.

Practice next

1. Load-test with k6 simulating 500 concurrent CRM streams.
2. Watch Postgres pool exhaustion and Redis latency.
3. Partition AiUsageDaily by month; archive old partitions.
4. Introduce CQRS read model for CRM agent facts.
5. Cap concurrent streams per tenant in middleware.

Remember

Scale API, cache, DB pool, and queues together. Load-test streams and RAG not just health endpoints. Partition large telemetry and usage tables.

Tenfold tenant growth

AgentNest grows from 50 to 500 tenants in one quarter.

Outcome: DbContext pool tuning and Redis session layer prevent connection storms.