n8n Cloud or self-hosting: what must your business organise?

Define the continuity your process needs

Ask what happens when the workflow is unavailable for an hour, a working day or longer. A weekly report has different requirements from a request route that staff depend on throughout the day. Name the largest backlog the team can handle manually and the information needed for that fallback. Availability then becomes an operational requirement rather than a vague adjective.

Distinguish recovery of the environment from recovery of business processing. A server may run again while some requests remain unfinished. Equally, the team may continue temporarily before the technical problem is solved. Give both routes an owner and an identifiable check. Calling a process business-critical becomes useful only when the organisation understands which interruption remains acceptable and what work must be preserved.

Compare managed and self-managed operation

n8n offers a managed cloud route and options for hosting it yourself. Self-hosting means choosing and maintaining an installation environment; with cloud use, the service model determines which operating activities the supplier performs. Examine the current agreement rather than assuming that every technical or process problem receives the same support from the same party.

Prepare a responsibility list covering installation, updates, monitoring, accounts, external connections, workflow rules and recovery of failed cases. Name a person or party for each activity and explain how handover works. A hosting provider might maintain infrastructure without understanding the meaning of your CRM fields. That business responsibility must remain somewhere else, even when the technical environment is maintained professionally and the underlying service is available.

n8n: Host n8n ↗

Draw the actual data journey

A self-hosted workflow may send information to a CRM, email service, AI model or logging facility. The location of n8n therefore does not identify where all processing occurs. For every connection, record what leaves, why it is needed and how long it remains necessary. A simple data-flow drawing is more informative than a collection of broad security assurances.

Include error notifications and test data. A notification containing a complete document may disclose more than normal processing does. Agree which reference is enough to locate a case and who may inspect its contents. Appropriate contractual assessment depends on the parties and information involved; a technical location decision does not answer those questions by itself. Check arrangements before real sensitive information is introduced into a trial.

A hosting decision also needs arrangements for updates, recovery and day-to-day ownership: Maintaining n8n workflows: ownership, checks and controlled change.

Restrict access by role and environment

Decide who may edit workflows, who manages connections and who only investigates errors. Developers do not automatically need all production information. Where practical, use separate test connections and representative example data. Describe how new staff receive access and how it is removed when an assignment or relationship ends. Ownership should remain clear even when a specialist performs administration.

Official security documentation lists several configuration options. Usability depends on edition, subscription and configuration. Verify required roles and security arrangements in the selected environment. A description of a feature does not prove that it has been correctly enabled for your accounts. Record a limited access exercise and the exceptions that remain, including situations where a required role must temporarily receive broader privileges under an agreed procedure.

n8n: Security ↗

Turn restoration into an actual exercise

A backup is useful only when the necessary components can be restored together. Inventory the database, workflow configuration, relevant attachments and the material needed to keep stored connections usable. Protect sensitive components and state who may retrieve them. Check the documentation for the chosen installation: losing one essential component can make an apparently complete backup unusable when recovery is needed.

Restore first into an isolated environment and prevent restored workflows from immediately sending real messages or changing external records. Check representative cases and reconcile technical execution history with business records. The exercise should reveal missing information and work requiring manual catch-up. Record observed duration and limitations. A planned restoration procedure is not equivalent to a successful recovery exercise performed under controlled conditions.

Assess load using the actual workload

Do not consider only daily trigger counts. Large attachments, simultaneous requests, long waits and extensive loops can impose very different loads on an environment. Describe ordinary volume and peaks separately. A trial should include the largest expected files and relevant concurrency without unnecessarily stressing production or connected services. Use evidence from that trial to revise capacity assumptions.

Choose the technical arrangement in response to findings. A larger architecture may help scale or continuity but introduces more components to operate. Ask what problem each additional component solves and who can investigate it during an outage. A simple installation that meets measured requirements and can be restored confidently may fit better than an elaborate arrangement that nobody has enough time or experience to support.

Keep licensing separate from hosting

Being able to install software does not mean every commercial use has identical terms. n8n distinguishes internal use, certain services and arrangements where others build or configure workflows. Cloud use has separate terms. Check your actual scenario against current official licence guidance and obtain supplier clarification when the intended arrangement is uncertain.

Describe who uses the editor, who can change logic and whether customers can configure their own automations. Resolve these characteristics before a product or service depends on the design. Do not justify the choice simply by calling the installation open source. Access to source code, operating an environment yourself and permission to offer a particular commercial application are separate questions that deserve separate answers.

n8n: Which license do I need for my use case? ↗

Work through a fictional choice

A fictional administration team wants to register enquiries daily and has no infrastructure operating team. Its main objective is a dependable first route with understandable alerts. A managed service may therefore be a sensible candidate, provided required features, data arrangements and external connections fit. The team remains responsible for process rules and for investigating unfinished requests that need a business decision.

A fictional software organisation instead has an established operations team, controlled infrastructure and demonstrated restoration routines. Self-hosting might fit that environment but still needs checks on licensing, access and load. These are possible decisions, not recommendations determined simply by company size. The decisive factor is the relationship between process requirements and people actually available, rather than a general preference for cloud services or owning a server.

Record the decision and an exit route

Keep an overview of the hosting choice, responsibilities, cost categories and accepted limitations. Describe the information and connections required for a later move. A workflow export may not include everything another environment needs. Plan how outstanding requests will finish and how duplicate execution is prevented while the old and new installations overlap during a controlled transition.

Ask proposals to distinguish hosting, technical administration and functional workflow maintenance. Establish incident reporting, any required out-of-hours availability and the boundary of additional work. The appropriate environment is one whose responsibilities make sense and whose everyday use, maintenance and recovery can actually be carried out by the parties involved. That is a stronger basis than selecting infrastructure from price or technical preference alone.

Turn the explanation into a project decision

Record your assumptions, unknowns and acceptance checks in the matching tool. Use the service guides to discuss delivery and scope.

• n8n consulting ↗
• Business process automation ↗

Questions about this topic

Continue with a related question