---
title: "Send from Cloudflare Workers"
description: "Export Workers logs and traces to EZGH Observability with Cloudflare's OTLP destinations, choosing the environment and application with headers."
---

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

# Send from Cloudflare Workers

Cloudflare Workers can export their logs and traces over OTLP to **destinations** you set up in
the Cloudflare dashboard. A destination sets an endpoint and custom headers, but not resource
attributes. Send the environment and application as the `x-ezgh-environment` and
`x-ezgh-application` headers instead (see [Service metadata](/observability/#service-metadata)).

## Create an ingest key

Create a key that can send both signals:

```sh
ezgh ingest-keys create cloudflare-workers --project web \
  --action obs.logs.ingest,obs.traces.ingest > ingest-key.txt
```

## Add the destinations

In the Cloudflare dashboard, go to **Observability** > **Destinations** and add one destination
for each signal:

| Field | Traces | Logs |
| --- | --- | --- |
| Destination name | `ezgh-traces` | `ezgh-logs` |
| Destination type | Traces | Logs |
| OTLP endpoint | `https://obs-ingest.ezghcloud.com/v1/traces` | `https://obs-ingest.ezghcloud.com/v1/logs` |

Give both destinations the same custom headers:

| Header | Value |
| --- | --- |
| `authorization` | `Bearer ezgh_ik_…` (your ingest key) |
| `x-ezgh-environment` | `production` |
| `x-ezgh-application` | `storefront` |

Workers sharing one environment and application can share these destinations. Create another
pair for each further environment, such as `staging`.

## Export from the Worker

Point the Worker's observability settings at the destinations, then redeploy it:

```jsonc
// wrangler.jsonc
{
  "observability": {
    "traces": { "enabled": true, "destinations": ["ezgh-traces"] },
    "logs": { "enabled": true, "destinations": ["ezgh-logs"] }
  }
}
```

Each Worker's records carry its own `service.name`, so you can tell several Workers apart within
one application.

## How the headers apply

- The headers fill in `ezgh.metadata.environment` and `ezgh.metadata.application` only on a
  resource that doesn't set them. When a resource sets its own value, that value is kept, even an
  invalid one (which is stored as `unknown`).
- Header values follow the same rules as the attributes: they're trimmed and lowercased, then
  must be 1 to 63 letters, digits, `.`, `_` or `-`, starting with a letter or digit.
- An invalid header value, or a header sent twice, rejects the whole request with `400`. Check
  the destination's headers if Cloudflare reports failed exports.
- Without the headers, Workers records go to the `unknown` environment.

Source: https://docs.ezghcloud.com/observability/send-from-cloudflare-workers/index.mdx
