FAQ
Common questions about how Portlane runs, deploys, and keeps your data on-prem
Data and hosting
Does my data ever leave my network?
No, by default. Portlane runs entirely on infrastructure you control — engines, Postgres, and the Control Tower console all live on your VM. Only the systems your flows explicitly target ever receive data.
Can I self-host on any cloud, or on-prem hardware?
Yes. Portlane ships as a Docker Compose bundle for a single Linux VM or bare metal. We publish sizing guidance for Hetzner Cloud, but any Linux host with Docker works.
Do I need to install anything on Windows or macOS machines?
No. Portlane's engines run as Linux containers on your server — there is no per-machine agent to install or maintain.
License and portal
How do I get my license file?
Licensed customers download it from the client portal. Sign in, open Licence, and download the file for your organization. The license reflects your commercial agreement — plan, seat count, enabled features, and validity period. Download is self-service once Bastion Lab has activated your organization. If no license appears, contact Bastion Lab.
Does Portlane phone home?
Your integration data stays on your instance by default. The supervision layer may optionally report coarse version and health metadata back to Portlane for support visibility — best-effort telemetry, not DRM, with no effect on how your instance runs. The data plane never transits through Portlane's cloud.
Flows and development
How do I build a flow — visually, or as code?
As code. A flow is a small declarative identity card (source, routing, target) plus a pure Python transform, both versioned in your own git repository and reviewed like any other change.
How does deployment work — do I push to the server?
No, the server only pulls. CI builds a versioned image and publishes routing changes; a self-hosted runner on your instance pulls the new image and applies the change. No inbound access to your server is ever required.
What happens if a message fails to process?
It's retried with backoff, then moved to a dead-letter state after the retry limit. Operators can inspect and replay it from the Control Tower — nothing is silently dropped.
How do rollbacks work?
Both the platform and your flow code are versioned container images. Rolling back means redeploying a previous, already-tested image — no manual reconfiguration.
Connectors and scope
What systems can Portlane connect to?
Databases, files, APIs, and streams out of the box, plus custom Python connectors for proprietary systems.
How is this different from a typical cloud iPaaS?
Most iPaaS platforms route your data through a vendor's cloud runtime. Portlane's engines run on your own infrastructure — the platform ships to you, not the other way around.
Can Portlane automate a web application that has no API?
That's the direction the ecosystem is extending toward: a browser automation agent that records a business user's actions once and replays them as a repeatable scenario, then delivers the resulting data to the platform like a connector would. It's aimed at legacy or third-party web apps that expose no API, database, or file interface — without the licensing weight of a general-purpose RPA suite.
Next steps
Still have questions? Book a demo.