Hosting decision guide
Best n8n hosting options depend on which work you want to own.
Hosting is not just where n8n runs. It decides who owns webhook URLs, SSL, volumes, Postgres, Redis, queue workers, backups, upgrades, and incident response.
Short Answer
Pick n8n Cloud if your main risk is operations. Pick a VPS with Docker Compose if you need self-hosted control and can prove persistence, backup restore, and public URL routing. Pick a managed automation alternative when you are building a new SaaS automation and do not need to preserve an n8n workflow model.
Hosting Fit Matrix
Self-hosted vs cloud| Option | Best for | Avoid when | Proof before production |
|---|---|---|---|
| n8n Cloud | Teams that want n8n without owning Docker, upgrades, SSL, backups, public webhook routing, or queue operations. | You require private network placement, custom infrastructure control, or a fully self-hosted compliance model. | A production webhook, credential refresh, and scheduled workflow run without you touching server config. |
| VPS with Docker Compose | Operators who want self-hosted control and can maintain volumes, Postgres, Redis, reverse proxy, SSL, and backups. | You do not have time to test backup restore, upgrades, DNS, and WEBHOOK_URL after every change. | A restart, backup restore, webhook smoke test, and upgrade rollback plan all pass. |
| Managed automation alternative | New SaaS workflows where the builder experience and managed platform matter more than n8n compatibility. | You need to repair an existing n8n instance or keep n8n-specific nodes and workflow exports. | The same trigger, data map, and output can be rebuilt with acceptable limits and pricing. |
Self-hosted readiness gates
- Workflows and credentials survive a container recreation.
- Public webhook URLs show the intended HTTPS domain.
- Backups include database, encryption key, binary data, and workflow exports.
- Postgres and Redis checks are documented if queue mode is enabled.
- Upgrade rollback is tested before the first serious production upgrade.
Managed path triggers
- Most incidents are SSL, DNS, Docker, queue workers, or backup operations.
- The workflow is business-critical but infrastructure ownership is not strategic.
- The team needs a reliable public webhook endpoint more than custom server control.
Recommended First Build
If you stay self-hosted, start with one production-like workflow, not a full migration. Test webhook delivery, one credential refresh, one scheduled run, one manual backup, one restore into a disposable instance, and one restart. If that sounds like too much, compare managed n8n or a platform rebuild before scaling the instance.