What causes "${componentType} template ${templateId} not found" in microsoft/autogen?

Inspect the chained cause (`except ValueError as e: print(e.__cause__)`) — the real Azure error is inside.

microsoft/autogen · error · Error

${componentType} template ${templateId} not found

Error message

${componentType} template ${templateId} not found

What it means

Catch-all raised when the Azure OpenAI embeddings.create call raises any exception — the original is chained via `from e`. Typical underlying causes are 401/403 auth errors, a deployment name that doesn't match `embedding_model`, wrong api_version for the endpoint, throttling (429), or network failures.

Source

Thrown at python/packages/autogen-studio/frontend/src/components/types/component-templates.ts:606

  componentType: ComponentTypes
): ComponentDropdownOption[] {
  const templates = getTemplatesForType(componentType);
  return templates.map((template) => ({
    key: template.id,
    label: template.label,
    description: template.description,
    templateId: template.id,
  }));
}

export function createComponentFromTemplateById(
  componentType: ComponentTypes,
  templateId: string,
  customLabel?: string
): Component<ComponentConfig> {
  const template = getTemplateById(componentType, templateId);
  if (!template) {
    throw new Error(`${componentType} template ${templateId} not found`);
  }

  return createComponentFromTemplate(templateId, componentType, {
    label: customLabel || `New ${template.label}`,
  });
}

// Specific helper functions for each component type
export function getTeamTemplatesForDropdown(): ComponentDropdownOption[] {
  return getTemplatesForDropdown("team");
}

export function createTeamFromTemplate(
  templateId: string,
  customLabel?: string
): Component<ComponentConfig> {
  return createComponentFromTemplateById("team", templateId, customLabel);
}

View on GitHub (pinned to 027ecf0a37)

Solutions

Inspect the chained cause (`except ValueError as e: print(e.__cause__)`) — the real Azure error is inside.
Set `embedding_model` to the exact deployment name on your Azure OpenAI resource.
Set `openai_api_version` to a version your resource supports (e.g. "2024-02-01" or newer).
For 429s add backoff/retry or a higher quota; for 401/403 fix keys/role assignment (Cognitive Services OpenAI User role).

Example fix

# before
config = AzureAISearchConfig(
    ..., embedding_provider="azure_openai",
    embedding_model="text-embedding-3-large",  # deployment mismatch
)

# after — model = your DEPLOYMENT name
config = AzureAISearchConfig(
    ..., embedding_provider="azure_openai",
    embedding_model="my-text-embedding-3-deployment",
    openai_api_version="2024-02-01",
)

Defensive patterns

Strategy: retry

Try / catch

import asyncio

async def embed_search(tool, query: str, attempts: int = 3):
    for i in range(attempts):
        try:
            return await tool.run(query)
        except ValueError as e:
            cause = e.__cause__
            status = getattr(cause, "status_code", None)
            if status == 429 and i < attempts - 1:
                await asyncio.sleep(2 ** i)
                continue
            raise

Prevention

Match embedding_model to the exact Azure deployment name.
Pin openai_api_version to one your resource supports.
Always inspect e.__cause__ before guessing at fixes.

When it happens

Trigger: Calling a vector search with embedding_provider='azure_openai' where `azure_client.embeddings.create(model=embedding_model, input=query)` throws: model not deployed on the resource, api_version mismatch (default '2023-11-01' vs resource expecting newer), invalid/insufficient credential, or transient 429/timeout.

Common situations: Using the base model name (text-embedding-ada-002) instead of the custom deployment name; defaulting to api_version 2023-11-01 against newer resources that require a supported version; hitting rate limits during load tests; key expired or rotated.

Related errors

AI-assisted analysis of microsoft/autogen@027ecf0a37 (2026-08-15). Data as JSON: /api/errors/7e876dcbf4084073. Report an issue: GitHub.