Performance Optimization

Performance Optimization

Guide to optimizing Reinhardt application performance.

Table of Contents

Routing Performance

matchit Radix Tree

Reinhardt uses matchit Radix Tree for O(m) route matching (m = path length).

// Fast matching even with 1000+ routes
let router = ServerRouter::new()
    .function("/api/resources/{id}", Method::GET, handler)
    .function("/api/resources/{id}/posts", Method::GET, handler)
    .function("/api/posts/{post_id}/comments/{comment_id}", Method::GET, handler);

Route Registration Best Practices

use reinhardt::ServerRouter;

// ✅ Use hierarchical routing
let api_router = ServerRouter::new()
    .with_prefix("/api/v1")
    .function("/users", Method::GET, list_users);

let router = ServerRouter::new()
    .mount("/api/", api_router);

// ❌ Register all routes flat (inefficient)
let router = ServerRouter::new()
    .function("/api/v1/users", Method::GET, list_users)
    .function("/api/v1/posts", Method::GET, list_posts);

Lazy Compilation

Routes are compiled on first access, so there's no startup overhead.

let router = ServerRouter::new()
    .function("/api/users", Method::GET, users_handler);

// Routes are compiled on first access
// Can explicitly compile with register_all_routes()
router.register_all_routes();

Database Query Optimization

Efficient Queries with Reinhardt's Query Builder

use reinhardt::query::prelude::{Query, Postgres, Expr, Func, Order};

// ✅ Query using indexes
let (limit, offset) = (page_size, (page - 1) * page_size);

let query = Query::select()
    .columns([User::Id, User::Username, User::Email])
    .from(User::Table)
    .and_where(Expr::col(User::IsActive).eq(true))
    .order_by(User::CreatedAt, Order::Desc)  // Sort by indexed column
    .limit(limit as u64)
    .offset(offset as u64)
    .to_owned();

// ✅ Count query is separate
let count_query = Query::select()
    .expr(Func::count(Expr::col(User::Id)))
    .from(User::Table)
    .and_where(Expr::col(User::IsActive).eq(true))
    .to_owned();

Avoid N+1 Problem

use reinhardt_db::orm::{NPlusOneConfig, NPlusOneScope};

// ❌ N+1 problem
let posts = Post::objects().all().await?;
for post in posts {
    let user = post.user().await?;  // Query in each loop
}

// ✅ Use a JOIN for a single-object relationship
let posts = Post::objects()
    .select_related(&["user"])
    .all()
    .await?;

// ✅ Use prefetching or a batch query for collections
let posts = Post::objects()
    .prefetch_related(&["comments"])
    .all()
    .await?;

// ✅ Make repeated query shapes visible in focused diagnostics
let (_, report) = NPlusOneScope::warn("posts.index", NPlusOneConfig::default())
    .run_with_report(async {
        render_posts().await
    })
    .await;

for finding in report.findings {
    tracing::warn!(
        fingerprint = %finding.fingerprint,
        count = finding.execution_count,
        "Repeated query shape detected"
    );
}

Use NPlusOneScope::fail(...).run(...) in focused tests when a path must stay batched. QuerySet execution and relationship accessors are recorded by active scopes. Use NPlusOneScope::spawn(...) for spawned tasks that should inherit the active scope. The detector is disabled unless a scope is installed, so production requests are unchanged by default.

Cache Strategies

HTTP Caching

HTTP caching using CacheControlMiddleware.

use reinhardt::staticfiles::{CacheControlMiddleware, CacheControlConfig};

let config = CacheControlConfig {
    public: true,
    max_age: 3600,        // 1 hour
    s_maxage: Some(86400), // 1 day on CDN
    immutable: true,       // Versioned files
};

let router = ServerRouter::new()
    .with_middleware(CacheControlMiddleware::new(config));

Cache Directives

use reinhardt::staticfiles::CacheDirective;

// Versioned files (cache for 1 year)
let immutable = CacheDirective::public()
    .with_max_age(31536000)
    .with_immutable(true);

// API responses (cache for 5 minutes)
let api_cache = CacheDirective::public()
    .with_max_age(300)
    .with_private(true);  // User-specific data

// HTML files (don't cache)
let no_cache = CacheDirective::no_cache();

ETag Conditional Requests

use reinhardt::ETagMiddleware;

let router = ServerRouter::new()
    .with_middleware(ETagMiddleware::new(ETagConfig::default()));

// When client sends If-None-Match header,
// returns 304 Not Modified if unchanged

Compression Settings

GZip Compression

use reinhardt::{GZipMiddleware, GZipConfig};

let config = GZipConfig {
    level: 6,  // Compression level (1-9, default 6)
    min_size: 1024,  // Minimum file size to compress
};

let router = ServerRouter::new()
    .with_middleware(GZipMiddleware::new(config));

Brotli Compression

use reinhardt::{BrotliMiddleware, BrotliConfig};

let config = BrotliConfig {
    quality: BrotliQuality::Medium,  // Compression quality
    min_size: 1024,
};

let router = ServerRouter::new()
    .with_middleware(BrotliMiddleware::new(config));

Choosing Compression Algorithm

FormatCompressionSpeedUse Case
GZipMediumFastGeneral text, JSON, HTML
BrotliHighMediumJavaScript, CSS, large static files

Compression Best Practices

// ✅ Provide both Brotli and GZip (browser chooses)
let router = ServerRouter::new()
    .with_middleware(BrotliMiddleware::new(brotli_config))
    .with_middleware(GZipMiddleware::new(gzip_config));

// Compressed response:
// Content-Encoding: br, gzip

Profiling

Measuring Request Time

use reinhardt::{MetricsMiddleware, MetricsConfig};
use std::time::Instant;

async fn timed_handler(req: Request) -> Response {
    let start = Instant::now();

    // Handler processing
    let response = process_request(req).await;

    let duration = start.elapsed();
    println!("Request took {:?}", duration);

    response
}

Metrics Collection

use reinhardt::{MetricsMiddleware, MetricsStore};

let store = MetricsStore::new();
let middleware = MetricsMiddleware::new(MetricsConfig::default());

let router = ServerRouter::new()
    .with_middleware(middleware);

See Also