# Railway Logs

Stream logs from Railway services into Gonzo with zero configuration. Railway's CLI outputs JSON over WebSocket. Gonzo auto-detects the format — no poller, no normalizer, no `jq` transforms.

### Quick Start

```bash
railway login
railway link
railway logs --json | gonzo
```

That's it. Works on all Railway plans (Trial, Hobby, Pro).

### Prerequisites

* Gonzo installed
* [Railway CLI](https://docs.railway.com/guides/cli) installed (`brew install railway`)
* A Railway project with at least one deployed service

### Log Types

| Log type       | CLI flag           | Gonzo compatible | What it captures                                 |
| -------------- | ------------------ | ---------------- | ------------------------------------------------ |
| **Deployment** | `--json` (default) | ✅                | App stdout/stderr                                |
| **Build**      | `--build --json`   | ✅                | Dependency installs, compilation, image creation |
| **HTTP**       | Dashboard only     | ❌                | Edge proxy request logs (not available via CLI)  |

### Usage Patterns

**Real-time streaming** (WebSocket, no rate limits):

```bash
railway logs --json | gonzo
```

**Fetch last N lines:**

```bash
railway logs --json --lines 100 | gonzo
```

**Time range:**

```bash
railway logs --json --since 1h | gonzo
```

**Build logs after a failed deploy:**

```bash
railway logs --build --json --lines 100 | gonzo
```

**Filter at source:**

```bash
railway logs --json --filter "@level:error" | gonzo
```

**Write to file:**

```bash
railway logs --json > /tmp/railway-logs.jsonl &
gonzo -f /tmp/railway-logs.jsonl --follow
```

**With local AI (logs never leave your machine):**

```bash
export OPENAI_API_KEY="ollama"
export OPENAI_API_BASE="http://localhost:11434"
railway logs --json | gonzo
```

### Multi-Service Projects

Stream from a specific service:

```bash
railway logs --json --service <service-id> | gonzo
```

### Rate Limits

**Streaming** (`--json | gonzo`) uses WebSocket — **not rate limited**. This is the recommended mode.

**Fetch mode** (`--lines`, `--since`) uses the GraphQL API:

| Plan       | Requests/hour |
| ---------- | ------------- |
| Free/Trial | 100           |
| Hobby      | 1,000         |
| Pro        | 10,000        |

**Platform cap:** 500 log lines/second per replica. Excess is silently dropped. Use minified JSON (not pretty-printed) to stay under the limit.

### Structured Logging Tips

If your app emits structured JSON to stdout, Railway preserves it as-is. Non-JSON is wrapped in `{"message":"...","level":"..."}`. Stderr output is auto-tagged as `level: "error"` — a common gotcha for Python's `logging` library which defaults to stderr.

**Time to complete:** 5 minutes **Prerequisites:** Railway CLI, Gonzo installed **Full guide:** [`guides/RAILWAY_USAGE_GUIDE.md`](https://github.com/control-theory/gonzo/blob/main/guides/RAILWAY_USAGE_GUIDE.md)


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.controltheory.com/controltheory-documentation/gonzo-docs/integration-examples/railway-logs.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.