Using Iii

<!-- generated by iii-skill-render. DO NOT EDIT (changes here are overwritten on the next render). Edit docs/next/using-iii/triggers.mdx. -->


{/* TODO: Re-link worker references to https://workers.iii.dev/workers/<name> once the Worker Docs migration ships. */}

<Note>
  Calling a function directly with `worker.trigger` or `iii trigger` are documented in [Using iii /
  Functions](./functions#triggering-invoking-functions). This page covers specifics about Triggers
  such as registering (binding) Triggers to Functions, gating triggers, and unregistering them.
</Note>

## Register a trigger

<Note>
  If you're authoring a worker, you'll want to refer to [Creating Workers /
  Triggers](../creating-workers/triggers#bind-a-function-to-an-existing-trigger-type) to learn the
  difference between registering a trigger, and registering a trigger type.
</Note>

Functions can also run when a trigger is satisfied. A trigger can be any event that happens such as
a request to an `http` endpoint, a `cron` job, a change in `state`, or any other trigger that a
worker supports. You can also [write your own](../creating-workers/triggers).

You bind triggers to functions via the `function_id`. The trigger declares its `type`, its `config`
(defined by each type), and the function to invoke.

<Tabs>
  <Tab title="Node / TypeScript">
    ```typescript
    import { registerWorker } from "iii-sdk";

    const url = process.env.III_URL;
    if (!url) throw new Error("III_URL must be set");
    const worker = registerWorker(url);

    worker.registerTrigger({
      type: "http",
      function_id: "math::add",
      config: { api_path: "/math/add", http_method: "POST" },
    });
    ```

  </Tab>
  <Tab title="Python">
    ```python
    import os
    from iii import register_worker, InitOptions

    worker = register_worker(
        os.environ.get("III_URL"),
        InitOptions(worker_name="my-worker"),
    )

    worker.register_trigger({
        "type": "http",
        "function_id": "math::add",
        "config": {"api_path": "/math/add", "http_method": "POST"},
    })
    ```

  </Tab>
  <Tab title="Rust">
    ```rust
    use iii_sdk::{InitOptions, RegisterTriggerInput, register_worker};
    use serde_json::json;

    let url = std::env::var("III_URL").expect("III_URL must be set");
    let worker = register_worker(&url, InitOptions::default());

    worker.register_trigger(RegisterTriggerInput {
        trigger_type: "http".into(),
        function_id: "math::add".into(),
        config: json!({ "api_path": "/math/add", "http_method": "POST" }),
        metadata: None,
    })?;
    ```

  </Tab>
</Tabs>

Per-type configuration is documented in each worker's Worker Docs (e.g.
[http](https://workers.iii.dev/workers/http) for the `http` type).

### Trigger metadata

The optional `metadata` field on a trigger registration (`null` / `None` in the examples above) is
arbitrary JSON stored with the trigger and delivered to the receiving function as a distinct
argument alongside the payload. It is useful for providing contextual information about the trigger
or execution context to the receiving function.

A target function shared by many triggers can use it to recover which registration fired and with
what context. See
[Receive per-invocation metadata](../creating-workers/functions#receive-per-invocation-metadata) for
handler-side access in each SDK.

<Note>
  `metadata` can be provided both via `registerTrigger` and direct `trigger()` invocations
</Note>

## Registering before a trigger type is available

Registration is order-independent. If you register a trigger whose type is not active in the project
yet (for example, the worker that publishes `http` has not connected), the engine stores the
registration optimistically and activates it automatically once that trigger type becomes available.
The engine logs a pending notice while a registration waits, and re-establishes stored registrations
when a publishing worker restarts, so a worker that comes up after its consumers receives trigger
registrations as expected.

Bind `math::add` to the `http` type before the http worker is running. The engine accepts the
registration and stores it:

<Tabs>
  <Tab title="Node / TypeScript">
    ```typescript
    // http worker not started
    worker.registerTrigger({
      type: "http",
      function_id: "math::add",
      config: { api_path: "/math/add", http_method: "POST" },
    });
    ```
  </Tab>
  <Tab title="Python">
    ```python
    # http worker not started
    worker.register_trigger({
        "type": "http",
        "function_id": "math::add",
        "config": {"api_path": "/math/add", "http_method": "POST"},
    })
    ```
  </Tab>
  <Tab title="Rust">
    ```rust
    use iii_sdk::RegisterTriggerInput;
    use serde_json::json;

    // http worker not started
    worker.register_trigger(RegisterTriggerInput {
        trigger_type: "http".into(),
        function_id: "math::add".into(),
        config: json!({ "api_path": "/math/add", "http_method": "POST" }),
        metadata: None,
    })?;
    ```
  </Tab>
</Tabs>

Start the http worker afterward. The engine activates the stored binding automatically, and the
endpoint serves requests without re-registering:

```bash
# http worker started after registration
iii worker add http

# the stored binding is now live on the http worker's default port (3111)
curl -X POST http://localhost:3111/math/add \
  -H 'Content-Type: application/json' \
  -d '{"a": 2, "b": 3}'
# 200 OK: the request reaches math::add through the activated trigger
```

For known trigger types (ex. `http`, `state`, `durable:subscriber`, `stream`), the pending notice
includes the install command for the worker that provides the type. For other types, find the worker
that exposes the needed type at [workers.iii.dev](https://workers.iii.dev).

A registration fails when an active provider rejects the config it was given (for example, an
invalid trigger config). The engine then sends a `TriggerRegistrationResult` with an `error` body
back to the worker that initiated the request and logs it.

## Bind multiple triggers to one function

It's valid to bind multiple triggers to the same `function_id` and this can be done across any
number of types. Register a second trigger with the same `function_id` and a different type or
config; the function runs unchanged whether the call arrives over HTTP, on a cron schedule, or from
a queue message.

<Tabs>
  <Tab title="Node / TypeScript">
    ```typescript
    // Same handler runs for an HTTP POST and a weekly cron tick.
    worker.registerTrigger({
      type: "http",
      function_id: "reports::generate",
      config: { api_path: "/reports/generate", http_method: "POST" },
    });

    worker.registerTrigger({
      type: "cron",
      function_id: "reports::generate",
      config: { expression: "0 0 9 * * 1" }, // Every Monday at 09:00
    });
    ```

  </Tab>
  <Tab title="Python">
    ```python
    worker.register_trigger({
        "type": "http",
        "function_id": "reports::generate",
        "config": {"api_path": "/reports/generate", "http_method": "POST"},
    })

    worker.register_trigger({
        "type": "cron",
        "function_id": "reports::generate",
        "config": {"expression": "0 0 9 * * 1"},  # Every Monday at 09:00
    })
    ```

  </Tab>
  <Tab title="Rust">
    ```rust
    use iii_sdk::RegisterTriggerInput;
    use serde_json::json;

    worker.register_trigger(RegisterTriggerInput {
        trigger_type: "http".into(),
        function_id: "reports::generate".into(),
        config: json!({ "api_path": "/reports/generate", "http_method": "POST" }),
        metadata: None,
    })?;

    worker.register_trigger(RegisterTriggerInput {
        trigger_type: "cron".into(),
        function_id: "reports::generate".into(),
        config: json!({ "expression": "0 0 9 * * 1" }), // Every Monday at 09:00
        metadata: None,
    })?;
    ```

  </Tab>
</Tabs>

## Gate a trigger with a condition

A trigger can carry an optional `condition_function_id` (set inside the trigger's `config`). When
the trigger fires, the engine invokes the condition function first with the same payload the handler
would receive; the target `function_id` only runs when the condition returns truthy. The condition
is a regular registered function.

<Tabs>
  <Tab title="Node / TypeScript">
    ```typescript
    worker.registerFunction(
      "orders::is-priority",
      async (payload: { customer_tier: string }) => payload.customer_tier === "gold",
    );

    worker.registerTrigger({
      type: "http",
      function_id: "orders::expedite",
      config: {
        api_path: "/orders/expedite",
        http_method: "POST",
        condition_function_id: "orders::is-priority",
      },
    });
    ```

  </Tab>
  <Tab title="Python">
    ```python
    def is_priority(payload: dict) -> bool:
        return payload.get("customer_tier") == "gold"

    worker.register_function("orders::is-priority", is_priority)

    worker.register_trigger({
        "type": "http",
        "function_id": "orders::expedite",
        "config": {
            "api_path": "/orders/expedite",
            "http_method": "POST",
            "condition_function_id": "orders::is-priority",
        },
    })
    ```

  </Tab>
  <Tab title="Rust">
    ```rust
    use iii_sdk::{RegisterFunction, RegisterTriggerInput};
    use schemars::JsonSchema;
    use serde::Deserialize;
    use serde_json::json;

    #[derive(Deserialize, JsonSchema)]
    struct Payload { customer_tier: String }

    worker.register_function(RegisterFunction::new(
        "orders::is-priority",
        |input: Payload| -> Result<bool, String> {
            Ok(input.customer_tier == "gold")
        },
    ));

    worker.register_trigger(RegisterTriggerInput {
        trigger_type: "http".into(),
        function_id: "orders::expedite".into(),
        config: json!({
            "api_path": "/orders/expedite",
            "http_method": "POST",
            "condition_function_id": "orders::is-priority",
        }),
        metadata: None,
    })?;
    ```

  </Tab>
</Tabs>

## Unregister a trigger

Trigger registration returns a handle with an `unregister()` method. Call it to drop the trigger at
runtime; when the worker disconnects, all of its triggers are removed automatically.

<Tabs>
  <Tab title="Node / TypeScript">
    ```typescript
    const trigger = worker.registerTrigger({
      type: "http",
      function_id: "math::add",
      config: { api_path: "/math/add", http_method: "POST" },
    });

    trigger.unregister();
    ```

  </Tab>
  <Tab title="Python">
    ```python
    trigger = worker.register_trigger({
        "type": "http",
        "function_id": "math::add",
        "config": {"api_path": "/math/add", "http_method": "POST"},
    })

    trigger.unregister()
    ```

  </Tab>
  <Tab title="Rust">
    ```rust
    use iii_sdk::RegisterTriggerInput;
    use serde_json::json;

    let trigger = worker.register_trigger(RegisterTriggerInput {
        trigger_type: "http".into(),
        function_id: "math::add".into(),
        config: json!({ "api_path": "/math/add", "http_method": "POST" }),
        metadata: None,
    })?;

    trigger.unregister();
    ```

  </Tab>
</Tabs>