Self-hosted n8n slowdowns are almost always configuration, not capacity. A step-by-step playbook: audit Postgres, tune execution-data pruning, move binary data out of the database, and VACUUM to reclaim disk — with measured before/after results.
Production runbook for n8n secrets: the shared encryption key across queue-mode workers, injecting secrets with $env, OAuth2 refresh pitfalls, a credential health-check workflow, and zero-downtime rotation.
Go beyond error workflows: enable n8n’s Prometheus metrics endpoint, scrape workers, build the four alerts that predict outages, and add a self-monitoring workflow that pings Slack.
A practical guide to n8n expressions and data mapping: the items model, $json vs $items, five reshaping patterns, expressions vs the Code node, and the pitfalls that break workflows at 3am.
A working n8n workflow that turns a keyword into a structured, SEO-ready content brief: it pulls live SERP data, extracts competitor heading structure, and drafts an intent-aware outline with Claude. Includes node configs, JSON, and real throughput numbers.
Put n8n workflows under Git with a CI/CD pipeline: export as JSON, structure a repo, promote dev to staging to prod, and roll back safely on the self-hosted Community Edition.
Three production patterns for reliable, high-volume API ingestion in n8n: cursor and offset pagination, batching with Loop Over Items, and Retry-After exponential backoff with real node configs and measured results.
Harden n8n webhooks for production: verify HMAC signatures on the raw body, enforce idempotency with Postgres, and respond fast with async queue processing.
Build a free, self-owned rank-tracking pipeline in n8n: pull Google Search Console data on a schedule, store positions in Postgres, diff snapshots, and alert Slack only on meaningful drops.
Stop copy-pasting node blocks. Learn when and how to extract n8n sub-workflows with the Execute Workflow node and trigger – typed input contracts, each vs once mode, error handling across the boundary, and measured results from a real refactor.
Build one reusable n8n error workflow that catches every failure, retries transient errors with capped backoff, alerts a human on permanent ones, and dead-letters unrecoverable payloads for replay.
Stop silent failures in production. A three-layer model for resilient n8n workflows: node-level retries, a global Error Trigger workflow, and a dead-letter queue you can replay.