Appearance

Wrangler Errors ​

AwaitStep uses wrangler internally to bundle and deploy Cloudflare Workers. Wrangler errors are surfaced in the deploy log.

wrangler: command not found ​

Cause: The wrangler CLI is not installed in the container.

Fix: Pull the latest container image:

bash
docker compose pull
docker compose up -d

If running AwaitStep from source, install wrangler:

bash
pnpm add -g wrangler
# or
npm install -g wrangler

Authentication error: You must be logged in ​

Cause: Wrangler is trying to use its own authentication instead of the API token supplied by AwaitStep. This can happen if there is a stale ~/.wrangler/config file inside the container.

Fix:

  1. Stop the container.
  2. Remove the wrangler config volume or exec into the container and delete ~/.wrangler/config.
  3. Restart the container.

If running from source:

bash
rm -rf ~/.wrangler/config

A worker with this name already exists ​

Cause: A Worker with the generated name already exists in your Cloudflare account, created outside of AwaitStep.

Fix: Either:

  • Delete the conflicting Worker in the Cloudflare dashboard.
  • Or rename your AwaitStep workflow to generate a different Worker name.

Worker names are derived from the workflow name and the APP_NAME env var.

Error: Script startup exceeded CPU time limit ​

Cause: The generated Worker script fails to initialise within Cloudflare's startup CPU time limit. This can happen if the workflow has very complex generated code.

Fix: Simplify the workflow. If the issue persists, file a bug — this may indicate a code generation issue.

Error: Too many resources bound ​

Cause: The workflow references too many KV namespaces, D1 databases, R2 buckets, or other Cloudflare bindings. Cloudflare limits the number of bindings per Worker.

Fix: Reduce the number of resources used by the workflow. Consider consolidating into fewer namespaces or databases.

TypeError: Cannot read properties of undefined in deploy log ​

Cause: A node's config has a missing required field that produced invalid generated code.

Fix:

  1. Open the workflow canvas.
  2. Look for nodes with a warning indicator.
  3. Open each node's config drawer and ensure all required fields are filled.
  4. Re-deploy.

Build failed: Module not found ​

Cause: The workflow or a node template imports a module that is not available. This can happen with marketplace nodes that have NPM dependencies not bundled correctly.

Fix:

  1. Check if the node's package dependencies are listed in its node.json dependencies field.
  2. If using a custom node, ensure all dependencies are declared.
  3. For marketplace nodes, try uninstalling and reinstalling the node.

Deployment succeeds but Worker behaves unexpectedly ​

Cause: The generated code is valid but has a logic error introduced by an expression or custom Step code.

Fix:

  1. Click View Generated Code in the workflow toolbar (or fetch the version via API).
  2. Review the generated TypeScript for obvious errors.
  3. Check that all {{expression}} references resolve to the correct values.
  4. Add console.log statements to Step nodes to debug (visible in Cloudflare Worker logs). Remove them before the final deploy.

Checking wrangler version ​

To check which version of wrangler AwaitStep is using:

bash
docker compose exec awaitstep wrangler --version

If you need a specific wrangler version, override the image or build from source with the desired version pinned.