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

# Actor logs

> Read a StatefulActor type's recent runtime and invocation logs with the CLI.

> **Requires CLI v0.5.8 or newer.**

`telnyx-edge actors logs <type>` prints a StatefulActor type's recent logs, oldest line first. It reads a window of history and exits — it does not stay attached. Without `--type`, both log streams print interleaved by time:

```bash theme={null}
telnyx-edge actors logs Counter
```

```text theme={null}
[2026-09-25T19:39:55Z] [log] incremented to 3
[2026-09-25T19:39:55Z] alice increment ok 2ms
```

## Runtime logs

**Runtime logs** are the `console.log`/`error` output your method bodies write — your actor's own logging, nothing else:

```bash theme={null}
telnyx-edge actors logs Counter --type runtime
```

Runtime lines print as `[timestamp] [level] message`.

## Invocation logs

**Invocation logs** are one platform-recorded record per method call — auto-RPC, `fetch()`, or `alarm()`: which instance handled it, the method, the outcome (`ok`/`error`), and how long it took. Invocation logs report traffic even for a type that logs nothing:

```bash theme={null}
telnyx-edge actors logs Counter --type invocations
telnyx-edge actors logs Counter --type invocations --instance alice   # one instance's records
```

Invocation lines print as `[timestamp] instance method outcome <duration>ms`.

Unlike a function, whose invocations are HTTP requests, an actor's invocations are method calls routed through the actor runtime — a function calling `env.COUNTER.increment(1)` produces one invocation record per call, attributed to the instance that handled it.

## Flags

| Flag | Description |
| - | - |
| `--type` | Log type to fetch: `runtime` or `invocations` (omit for both, interleaved) |
| `--instance` | Keep only this instance's records; valid only with `--type invocations` |
| `--since` | How far back to read (e.g. `10m`, `2h`; default `1h`, max `24h`) |
| `-n`, `--last` | Maximum lines to print (default `50`, max `250`) |
| `--json` | Output the result as a JSON object instead of formatted lines |

`--since` and `--last` are clamped rather than refused when they exceed the maximum. There is no `--tail` for actor logs yet — rerun the command to read a fresh window; the function-level `--tail` does not apply to actor types.

Full command reference: [CLI reference](/docs/edge-compute/reference/cli#actors-logs).