Support
A deployment you control needs support that understands your environment, not a queue that assumes it can log in and look. Here is how we work, what we commit to, and how to reach us.
Getting help
Start with the one that matches the urgency. Everything is tracked, and nothing requires you to give us access to your data.
The main route for questions, configuration help, connector issues and feature requests. Sign in to the support portal to raise and track tickets, or email us directly and we will create one for you.
Open the support portal →For production outages and retrieval failures affecting live users. Incidents are escalated immediately rather than queued behind standard tickets. Suspected security issues go to the security channel instead, so they reach the right team first time.
[email protected] →A dedicated channel for security researchers and customer security teams, with a responsible disclosure policy and a one working day acknowledgement.
Read the policy →Never paste customer documents, personal data or credentials into a support ticket. Zegress support does not need your data to diagnose a problem. Logs, configuration and error messages are enough, and we will tell you exactly what to redact if we ask for anything more.
Service levels
Response targets are measured in business hours from the point a ticket is raised, and apply to all customers on a current licence.
| Severity | What it means | First response | Update cadence |
|---|---|---|---|
| S1 · Critical | Production is down, or a security issue is suspected. No workaround exists and users are blocked. | 1 business hour | Every 2 hours until resolved |
| S2 · High | Major function degraded, such as indexing or retrieval failing on a connected source. A workaround may exist. | 4 business hours | Daily |
| S3 · Normal | A defect or question that does not block production use. Configuration help sits here. | 1 business day | Every 3 business days |
| S4 · Low | Feature requests, documentation gaps and general guidance. | 3 business days | On change |
Business hours are 09:00 to 17:30 UK time, Monday to Friday, excluding English public holidays. Enterprise agreements can extend cover to follow-the-sun and add named contacts with direct escalation. Severity is set by you when you raise the ticket; we will only change it after agreeing the change with you.
Support plans
Enhanced and Sovereign are agreed as part of an Enterprise private offer. Talk to us about which one fits your environment.
Raising a good ticket
Support for self-hosted software depends on what you can tell us. These four things resolve most tickets on the first response.
Which version you are running, whether it is connected or air-gapped, and which model you have K-Lake pointed at. This alone rules out most of the possibility space.
The question asked, the answer expected, and the answer received. If retrieval is the issue, tell us whether the source document is indexed and whether the user can open it directly.
Application logs, connector logs and error output. Redact anything sensitive first. We will never ask you to send source documents, and you should refuse if anyone does.
Everything marked critical means nothing is. Accurate severity is what lets us put the right person on your problem quickly.
Before you raise a ticket
Architecture, controls, certifications and documentation are published in the trust centre. Open the trust centre
Connectors, permissions, models and deployment options. Platform capabilities
Data residency, migration, air-gapped operation and billing, answered in full. Read the FAQ
Marketplace billing, plan changes and private offers are handled through your Azure account and your Zegress contact. Email us
Not a customer yet
Ask us anything about the architecture, the deployment or the security review. We would rather answer it now than have it surface in procurement.
Ask a question Or email [email protected] Try it on your data