Performance

Elasticsearch is a powerful search and analytics engine, but optimizing it for production requires understanding indexing strategies, query patterns, and cluster configuration. This guide covers essential optimization techniques. Cluster architecture Node roles Configure nodes with specific roles: # Master node node.roles: [master] # Data node node.roles: [data] # Ingest node node.roles: [ingest] # Coordinating node (default) node.roles: [] # No specific role Shard strategy Primary shards: Set at index creation (cannot be changed) ...

Low-effort wins php artisan config:cache, route:cache, view:cache; warm on deploy. Enable OPcache with sane limits; preloading for hot classes when applicable. Use queues for emails/webhooks; keep HTTP requests lean. Database hygiene Add missing indexes; avoid N+1 with eager loading; paginate large lists. Use read replicas where safe; cap per-request query count in logs. Measure slow queries; set alarms on p95 query time. HTTP layer Cache responses with tags (Redis) for fast invalidation. Use CDN for static/media; compress and set cache headers. Leverage middleware to short-circuit authenticated user cache when possible. Observability Laravel Telescope or Horizon for queues; metrics on throughput, failures, latency. Log DB/query counts; track opcache hit rate and memory usage. Checklist Config/route/view cached on deploy. OPcache enabled and monitored. DB queries optimized and indexed; N+1 checks in CI. Responses cached where safe; queues handle slow work.

Keys & TTLs Namespaced keys: app:domain:entity:id. Set TTLs per data volatility; use jitter to avoid thundering expirations. Version keys on schema changes to prevent stale reads. Stampede protection Use SETNX/lock around rebuild; short lock TTL with fallback. Serve stale-while-revalidate: return cached value, refresh asynchronously. Serialization & size Prefer JSON with bounded fields; compress only large blobs. Avoid massive lists/hashes; paginate or split keys. Operations Monitor hit rate, command latency, memory, evictions. Use connection pooling; set timeouts and retries with backoff. Cluster/replica for HA; read from replicas if consistency allows. Checklist Keys versioned; TTLs with jitter. Stampede controls in place. Metrics for hit/miss/latency/evictions; alerts configured.

Query hygiene Add composite indexes matching filters/order; avoid leading wildcards. Use EXPLAIN to verify index usage; watch for filesort/temp tables. Prefer keyset pagination over OFFSET for large tables. Config basics Set innodb_buffer_pool_size (50-70% RAM), innodb_log_file_size, innodb_flush_log_at_trx_commit=1 (durable) or 2 (faster). max_connections aligned with app pool size; avoid connection storms. Enable slow query log with sane threshold; sample for tuning. App considerations Reuse connections (pooling); avoid long transactions. Limit selected columns; cap payload sizes; avoid large unbounded IN lists. For read-heavy workloads, add replicas; route reads carefully. Checklist Indexes audited; EXPLAIN reviewed. Buffer pool sized; slow log enabled. Pagination and payloads bounded; connections pooled.

Pooling Use a pool (pg/pgbouncer); size = (CPU * 2–4) per app instance; avoid per-request connections. For PgBouncer in transaction mode, avoid session features (temp tables, session prep statements). Query hygiene Parameterize queries; prevent plan cache thrash; set statement timeout. Add indexes; avoid wild % patterns; paginate with keyset when possible. Monitor slow queries; cap max rows returned; avoid huge JSON blobs. App settings Set statement_timeout, idle_in_transaction_session_timeout. Use prepared statements judiciously; for PgBouncer, prefer server-prepared off or use pgbouncer session mode. Pool instrumentation: queue wait time, checkout latency, timeouts. OS/DB basics Keep Postgres on same AZ/region; latency kills. Tune work_mem, shared_buffers, effective_cache_size appropriately (DB side). Use EXPLAIN (ANALYZE, BUFFERS) in staging for heavy queries. Checklist Pool sized and monitored; PgBouncer for many short connections. Query timeouts set; slow logs monitored. Key indexes present; pagination optimized. App-level metrics for pool wait, query latency, error rates.

This note condenses the DEV article “Mastering Frontend Performance: Harnessing the Power of Caching” into actionable steps for modern apps. Why cache Reduce network and CPU cost for repeated data/computation. Improve perceived speed and resilience to flaky networks. Keep UIs responsive under load. Layers to combine HTTP caching: set Cache-Control, ETag, Last-Modified, stale-while-revalidate for API/static responses; prefer immutable, versioned assets. Client memoization: cache expensive computations/render data (useMemo, useCallback, memoized selectors). Data caching: use React Query/SWR/Apollo to dedupe fetches, retry, refetch on focus. Service worker (when appropriate): offline/near-edge caching for shell + static assets. React hook hygiene Memoize derived data: useMemo(() => heavyCompute(input), [input]). Memoize callbacks passed to children to avoid re-renders: useCallback(fn, deps). Keep props stable; avoid recreating objects/functions each render. HTTP cache playbook Static assets: long max-age + immutable on versioned filenames. APIs: choose strategy per route: idempotent reads: max-age/stale-while-revalidate with ETag. personalized or sensitive data: no-store. list endpoints: shorter max-age + revalidation. Prefer CDN edge caching; compress (Brotli) and serve modern formats (AVIF/WebP). UI checks No spinner longer than a couple of seconds; use skeletons and optimistic updates where safe. Avoid layout shift when cached data arrives—reserve space. Track Core Web Vitals (LCP/INP/CLS) and hit-rate for key caches. Quick checklist Versioned static assets + long-lived caching headers. API cache policy per route with ETag/stale-while-revalidate. React memoization for heavy work and stable callbacks. Data-layer cache (React Query/SWR) with sensible stale times + retries. RUM/CI dashboards watching Web Vitals and cache hit rates. Takeaway: Combine HTTP caching, client memoization, and data-layer caches to ship faster pages and keep them fast under real traffic.

Deadlines & retries Require client deadlines; enforce server-side context with grpc.DeadlineExceeded handling. Configure retry/backoff on idempotent calls; avoid retry storms with jitter + max attempts. Interceptors Unary/stream interceptors for auth, metrics (Prometheus), logging, and panic recovery. Use per-RPC circuit breakers and rate limits for critical dependencies. TLS & auth Enable TLS everywhere; prefer mTLS for internal services. Rotate certs automatically; watch expiry metrics. Add authz checks in interceptors; propagate identity via metadata. Resource protection Limit concurrent streams and max message sizes. Bounded worker pools for handlers performing heavy work. Tune keepalive to detect dead peers without flapping. Observability Metrics: latency, error codes, message sizes, active streams, retries. Traces: annotate methods, peer info, attempt counts; sample smartly. Logs: structured fields for method, code, duration, peer. Checklist Deadlines required; retries only for idempotent calls with backoff. Interceptors for auth/metrics/logging/recovery. TLS/mTLS enabled; cert rotation automated. Concurrency and message limits set; keepalive tuned.

Process manager modes pm=dynamic for most apps; pm=static only when workload is predictable and memory bounded. Key knobs: pm.max_children, pm.start_servers, pm.min_spare_servers, pm.max_spare_servers. Size max_children = (available RAM - OS/webserver/DB) / avg worker RSS. Opcache Enable: opcache.enable=1, opcache.enable_cli=0, opcache.memory_consumption sized for codebase, opcache.interned_strings_buffer, opcache.max_accelerated_files. Avoid opcache.revalidate_freq=0 in prod unless you control deploy restarts; prefer deploy-triggered reloads. Timeouts & queues Keep request_terminate_timeout sane (e.g., 30s-60s); long requests move to queues. Use pm.max_requests to recycle leaky workers (e.g., 500-2000). Watch slowlog to catch blocking I/O or heavy CPU. Observability Export status_path and scrape: active/idle/slow requests, max children reached. Correlate with Nginx/Apache logs for upstream latency and 502/504s. Alert on max children reached, slowlog entries, and rising worker RSS. Checklist Pool sizing validated under load test. Opcache enabled and sized; reload on deploy. Timeouts/queues tuned; slowlog on. Status endpoint protected and scraped.

The N+1 problem is a common performance issue in JPA. Here’s how to solve it in Spring Data JPA. Understanding N+1 Problem // N+1 queries: 1 for users + N for each user's posts List<User> users = userRepository.findAll(); for (User user : users) { List<Post> posts = postRepository.findByUserId(user.getId()); // N queries } Solutions 1. Eager Fetching @Entity public class User { @OneToMany(fetch = FetchType.EAGER) private List<Post> posts; } 2. Join Fetch (JPQL) @Query("SELECT u FROM User u JOIN FETCH u.posts") List<User> findAllWithPosts(); 3. Entity Graphs @EntityGraph(attributePaths = {"posts"}) @Query("SELECT u FROM User u") List<User> findAllWithPosts(); 4. Batch Fetching @Entity public class User { @BatchSize(size = 10) @OneToMany(fetch = FetchType.LAZY) private List<Post> posts; } Best Practices Use lazy loading by default Fetch associations when needed Use entity graphs Monitor query performance Use DTO projections Conclusion Solve N+1 problems with: ...

This article summarizes the DEV post “Key Performance Metrics for Web Apps and How to Measure Them,” focusing on the most important signals and how to capture them. Core metrics LCP (Largest Contentful Paint): loading speed; target < 2.5s p75. INP/TTI (Interaction to Next Paint / Time to Interactive): interactivity; target INP < 200ms p75. FCP (First Contentful Paint): first visual response. TTFB (Time to First Byte): server responsiveness. Bundle size & request count: total transferred bytes and requests. Measurement toolbox Lighthouse: automated audits; budgets and suggestions. WebPageTest: multi-location runs, filmstrips, waterfalls. RUM (GA or custom): real-user timing for live traffic. React profiler & perf tools: find slow renders/update frequency. Webpack/Vite bundle analyzer: visualize bundle composition and dead weight. Optimization reminders Ship less JS/CSS; enable tree shaking/code splitting. Compress and cache static assets; serve modern image formats. Trim request count; inline critical CSS for hero; defer non-critical JS. Watch layout stability; reserve space to avoid CLS hits. Set and enforce budgets (JS gz < 200KB, LCP < 2.5s p75, INP < 200ms). Takeaway: Track a small, high-signal set of metrics with both lab (Lighthouse/WPT) and field (RUM) data, then enforce budgets so regressions fail fast.