I would like to conduct a reliability demonstration test for a product to ensure that the product will meet lifetime requirement. I would like to keep the sample size small (around 10) and test it until failure for sever user. The product is a consumer product (vacuum cleaner). I would like to understand which is better? 'having a 90% reliable product at 50% confidence interval' OR 'having a 50% reliable product at 90% confidence interval'. Is there any recommended practice in the industry?

thanks