n8n Cloud or self-hosting? Rehearse a recovery before choosing
Cloud fits teams without someone to run the platform. If workflows must run in your own environment, prove that a restore can reconnect systems and process missed work before choosing self-hosting.
n8n connects applications, data, and business rules through nodes. Use managed n8n Cloud or run it on your own infrastructure. If your team connects spreadsheets and ticketing systems but has nobody on call for servers, start with Cloud. Consider self-hosting when the workflow must run in your own environment and someone owns upgrades, backups, and recovery. Compare who can get the workflow running again, not just who can build it.
Recover the business process, not just the canvas
Say you classify new form submissions as pending or assigned every morning, then sync them to a help desk. Use synthetic data to rehearse an outage. Bring the workflow into an isolated test instance, connect test credentials, process two sample records that were missed, and check that replay doesn’t open duplicate tickets. n8n documents importing exported workflows; its downloadable Cloud workflow backups don’t include credentials. A configuration file alone won’t reconnect your systems. On Cloud too, assign someone to rebuild connections, identify missing tickets, and verify replay results.
Have a second person rehearse self-hosted recovery
The self-hosting docs say an encryption key protects credentials before they’re saved to the database. Instance recovery needs a controlled way to restore both the database and its original credential key. Don’t count a restarted container as a successful recovery. Ask a second authorized maintainer to restore in a test environment and verify that the workflow actually completes its task. Budget for routine upgrades, backup checks, drills, and incident coverage. If nobody can own that work and Cloud meets your data and connectivity requirements, choose Cloud. If self-hosting is mandatory but the pilot hasn’t passed, keep the existing process. Don’t put a critical workflow on an instance whose recovery you haven’t rehearsed.
Content note:This article is based on publicly available information and is intended as a reference for evaluating tools. Features, pricing, and related policies may change at any time; please refer to the latest official information.
Original source:blog.n8n.io