Opening a Support Ticket
- Go to https://portal.nebul.com/. The Help Center opens.
- Click Open a support ticket under Contact us.
- Fill in the required fields - see the table below for what actually helps us move fast.
- Attach logs, screenshots, or config exports. The more context, the faster we can help.
- Click Submit ticket. You'll receive a case number by email - keep it handy.
What to Include
| Field | What to include |
|---|---|
| Product | Select the relevant product: Inference API, PrivateGPT, Private NeoCloud (Compute / Kubernetes / Storage / Managed Databases), or 'Other'. |
| Affected Resource | GPU cluster or node type (e.g. DGX H100, HGX H200), Kubernetes cluster, inference endpoint URL, instance ID, or storage volume - whichever applies. |
| Priority | Use the priority table below. If unsure, start at P3 - we'll adjust if needed. |
| Subject | One clear sentence: what and where. E.g. 'Inference API returning 429 on Qwen3 endpoint' or 'DGX cluster pods OOM-killed after scaling event'. |
| Description | Context and timeline, expected vs. actual behaviour, and full error output. For Inference API: include the model ID, endpoint (e.g. /v1/chat/completions), and full error response body. For NeoCloud: include GPU type, container or cluster config, and kubectl or OpenStack logs where available. |
Priority Levels
Set the right priority - it determines how fast we respond and who picks it up. Don't understate a P1 to be polite, and don't overstate a P4 hoping for faster service. We'll reclassify if needed, no penalty.
| Level | Description | S1 Response | Example |
|---|---|---|---|
| P1 | Service down | < 4h / < 1h / < 15 min | Production infra down, no workaround |
| P2 | Partial outage | Best effort same day | Critical errors, partial functionality, no workaround |
| P3 | Degraded service | Next business day | Service slow or degraded; workaround exists |
| P4 | Minor issue / question | Within 3 business days | Minor issue, no business impact, or general questions |