JuliusBrussee/caveman · error

cachebench: invalid replay model

Error message

cachebench: invalid replay model

What it means

Error "cachebench: invalid replay model" thrown in JuliusBrussee/caveman.

Source

Thrown at cacheengine/cachebench/replay_http.go:167

		return ReplayResponse{}, errors.New("cachebench: provider request failed")
	}
	defer response.Body.Close()
	body, err := io.ReadAll(io.LimitReader(response.Body, transport.maxResponseBytes+1))
	if err != nil {
		return ReplayResponse{StatusCode: response.StatusCode}, errors.New("cachebench: provider response read failed")
	}
	if int64(len(body)) > transport.maxResponseBytes {
		return ReplayResponse{StatusCode: response.StatusCode}, errors.New("cachebench: provider response exceeds configured limit")
	}
	return ReplayResponse{
		StatusCode: response.StatusCode, Body: body,
		ProviderRequestID: providerRequestID(response.Header),
	}, nil
}

func (transport *HTTPReplayTransport) endpoint(provider string, outbound ReplayOutbound) (*url.URL, error) {
	if !validBoundedText(outbound.Model, 512, false) {
		return nil, errors.New("cachebench: invalid replay model")
	}
	base := transport.baseURLs[provider]
	if provider == "bedrock" && base == nil {
		if !validAWSRegion(outbound.Region) {
			return nil, errors.New("cachebench: invalid Bedrock region")
		}
		parsed, _ := url.Parse("https://bedrock-runtime." + outbound.Region + ".amazonaws.com")
		base = parsed
	}
	if base == nil {
		return nil, fmt.Errorf("cachebench: unsupported replay provider %q", provider)
	}
	var path, rawPath string
	switch provider {
	case "openai":
		if outbound.Endpoint != "/v1/chat/completions" && outbound.Endpoint != "/v1/responses" {
			return nil, errors.New("cachebench: unsupported OpenAI replay endpoint")
		}

View on GitHub (pinned to 27d5a3981a)

Solutions

1. Use a valid model for the replay provider.

When it happens

Trigger: Thrown at cacheengine/cachebench/replay_http.go:167 when the library encounters an invalid state.

Common situations: See trigger scenarios.

AI-assisted analysis of JuliusBrussee/caveman@27d5a3981a (2026-08-15). Data as JSON: /api/errors/9633a59d637af52d. Report an issue: GitHub.