Request Hedging Tail Latency: Cutting p99 Without Overload
Learn how hedged requests cut p99 tail latency in microservices: delay tuning, hedge budgets, cancellation, idempotency, routing, and when to avoid it.
13 min read
Archive
All articles
295 published
Profiling a Slow Endpoint: Flame Graph to Fix, On-CPU vs Off-CPU
Learn to diagnose a slow monolith endpoint: scope the problem, use APM traces, pick on-CPU or wall-clock profiling, read flame graphs, and fix N+1 queries.
16 min read
nx migrate upgrade: A Two-Phase Guide to Nx Breaking Changes
Run an nx migrate upgrade as a reviewable change: inspect migrations.json, apply in small commits, verify behavior, and handle breaking changes safely.
16 min read
Nx Continuous Tasks: Run Dev Servers and APIs Together Locally
Mark serve targets as continuous in Nx and wire them with task dependencies so nx serve web starts your API too, plus readiness, ports and e2e.
13 min read
Nx Affected Inputs for Non-JS Files: SQL, Protobuf and Config Assets
Configure Nx named inputs, implicit dependencies and workspace-root filesets so changes to SQL, protobuf, YAML and template files trigger the right tasks.
13 min read
Nx Inputs for Non-JS Files: SQL, Protos and Configs in Affected
Model SQL migrations, .proto schemas and config files with Nx namedInputs and dependency edges so affected detection and caching stop skipping changed projects.
13 min read
How to Reduce NAT Gateway Costs: Attribute Bytes Before You Fix
Learn how NAT Gateway billing stacks with AWS data transfer costs, how to attribute bytes with flow logs, and when VPC endpoints actually save money.
13 min read
Undici Connection Pooling for Node.js Fetch: Sizing and Timeouts
How Node's fetch pools connections through undici, and how to size, scope, time out and observe dispatchers so outbound HTTP calls stop failing intermittently.
14 min read
Speed Up Monolith Startup with Lazy Module Loading: A Practical Guide
Learn how to profile application boot time, defer modules safely, keep readiness honest, and weigh lazy loading against first-request latency and late failures.
14 min read
Cluster Autoscaler vs Karpenter: How Node Scaling Really Differs
Compare Cluster Autoscaler and Karpenter by architecture, scale-down, spot handling and portability, with a decision guide and a safe migration path for EKS.
14 min read
In-Process Event Bus in a Monolith: Domain Events Done Right
Design an in-process event bus for a modular monolith: domain events, in-transaction vs after-commit handlers, outbox durability, boundaries and testing.
14 min read
Docker Multi-Stage Build for Python APIs: Venv-First Guide
Build smaller Python API images with a docker multi-stage build: pass one venv between stages, order layers for caching, use uv and slim bases safely.
13 min read