> ## Documentation Index
> Fetch the complete documentation index at: https://docs.buildpixel.io/llms.txt
> Use this file to discover all available pages before exploring further.

# Troubleshooting

> Symptoms you might see, what they mean, and how to fix them.

## Build fails

### Compile error in your code

```
LogPlayLevel: Error: MyGameModule.cpp(42): error C2065: 'undefined': undeclared identifier
```

A real C++ compile error. BuildPixel didn't break this — fix the file locally, push.

### Link error only on cloud, works locally

Most common cause: a plugin installed locally (e.g. via Epic launcher) but not in your repo. Check your `.uproject`'s `Plugins` array — anything you reference must be either in the engine, in `Plugins/` in your repo, or a known marketplace plugin.

### Cook failed on a specific asset

```
LogCook: Error: Failed to cook /Game/Maps/Level3
```

The asset is broken. Open in editor locally, fix or remove the broken dependency, commit. Cook is content-faulty, not code-faulty.

### Source sync fails — GitHub

```
fatal: could not read Username for 'https://github.com'
```

The GitHub App lost access to the repo. Check `github.com/settings/installations` → BuildPixel → repo selection.

### Source sync fails — Perforce

```
Perforce password (P4PASSWD) invalid or unset.
```

Ticket expired or password rotated. Update org's Perforce config in **Settings → Integrations**.

## Build never starts

If a build sits in `pending` for a long time:

1. **Check plan limits** — Free plan has concurrency 1. If a build is running, the next one waits.
2. **Try cancelling and re-triggering** — occasionally a queue gets wedged.

If `pending` for over an hour, contact support with the `buildId`.

## Dashboard issues

### Logs don't stream live

Check if the WebSocket connection is open in the dashboard's network tab. If it's blocked (corporate proxies sometimes do this), the dashboard falls back to polling.

### Logs show, then suddenly stop mid-build

The build process probably failed unrecoverably. The status will flip to `failed` within \~30 seconds.

## Auto-deploy

### Auto-deploy keeps failing

Check the build's `autoDeployStatus` and the error message:

* `503` from your endpoint → upstream busy; we retry. Make your endpoint more resilient.
* `401` → your endpoint rejected the signature. Verify your signature validation logic against [the webhook spec](/api-reference/webhooks/overview#verifying-signatures).
* `connection refused` → your endpoint isn't reachable from the public internet.

## Authentication

### `401 invalid_token`

* Token was revoked, expired, or mistyped.
* Check `/auth/me` to verify the token is good.

### Sudden 401 on a token that worked yesterday

* Did you accidentally commit it to git? We auto-revoke leaked tokens.
* Was the user that owns the token removed from the org? Personal tokens go away with the user.

## When you really need help

Email [support@streampixel.io](mailto:support@streampixel.io) with:

* **What you tried**
* **What happened** (screenshots, error messages)
* **`requestId`** from API errors, **`buildId`** for build issues

For Enterprise: your dedicated channel + `[P1]` in the subject for production issues.

We've seen most of the common failure modes — usually a 5-minute fix once we know what's going on.