It's an easy to use, no code solution which makes it very flexible when deploying various models across different environments (private/public clouds).
We've been using this product since 2016 but we're still finding ways of improving its performance by optimizing our workloads as well tuning up some parameters that might have caused us problems at first time such as memory allocation or IOPS limitations etc.. Also there is not much support from their side so you need your own resources! - Supportive community- Easy integration with other tools like AWS Kinesis Firehose.- Good monitoring dashboard provided out of box without needing additional plugins / addons installed within Jupyter Notebooks environment-.