Security you can verify.
octoja has far-reaching access to the devices you look after. That is why we set out openly how the platform is operated, which protective mechanisms are anchored in the product — and how to reach us if something strikes you as wrong.
Where the platform runs
Details of the infrastructure we operate ourselves.
Hetzner in Germany
The platform runs in a Kubernetes cluster at Hetzner. Their data centres are certified to ISO/IEC 27001 and attested under BSI C5. In normal operation, customer data does not leave Germany.
Recovery is rehearsed
Backups are not just created, they are restored regularly. A backup that has never been tested is not a backup.
Round-the-clock monitoring
The platform is monitored 24/7. You can check the current operational state yourself at any time.
Eight points you can hold us to
Properties of the software and fixed rules in our development process — several of them safeguarded by automated tests.
The agent connects outbound
The connection is always established from the endpoint to your instance, never the other way round, and runs exclusively over TLS. You do not have to open an inbound port or create a firewall rule for remote support.
Updates are verified and can be rolled back
Every downloaded file is verified via SHA256, and path traversal is blocked. The version switch is atomic and rolls back automatically if the start fails. Windows binaries are shipped signed.
No endpoint without a deliberate access decision
An architecture test fails the build as soon as an interface is neither explicitly protected nor explicitly opened. Forgetting is therefore technically impossible.
Login – state of the art
Single sign-on via Microsoft Entra ID, optionally with two-factor confirmation. Passwords are hashed with BCrypt, and the session cookie is exclusively HttpOnly and Secure. Revoked permissions take effect immediately — not at the next login.
Complete tenant isolation
Every customer receives their own environment with their own database — no shared data with other octoja customers. Within your environment, the permission model separates your own customers from one another, and does so in the database query itself rather than afterwards in memory.
Minimal data sent outward
What your instance reports back to us is limited to the instance identifier, the date and billing data. No hostnames, no IP addresses, no inventory data.
Four-eyes principle in code review
No change goes into the platform without a second person having reviewed and approved it. That applies to features just as much as to bug fixes.
Audits and penetration tests
The code is audited regularly. In addition, we test continuously in-house and commission external agencies to carry out penetration tests.
Have you found a vulnerability?
Then we would like to hear about it before anyone else does. We acknowledge receipt within two working days.
What we want to hear about — and what we do not
So that your report reaches the right people quickly and you remain on safe legal ground.
In scope
Please refrain from
If you stay within this framework, act in good faith and give us reasonable time to remediate, we will not treat your report as grounds for legal action. Please publish details only after agreeing them with us.
What happens after your report
Receipt acknowledged
You receive a response within two working days, naming a point of contact.
Assessment
We reproduce the finding, assess the impact and come back to you with an evaluation.
Remediation and feedback
We let you know as soon as a fix has shipped — and credit you as the finder if you wish.
Report a vulnerability
For confidential or encrypted communication, please use the email address directly. Everything else can be entered here.