Seldon

Maintenance release

* Kubernetes default memory requests and limits set for core services
* GlusterFS persistent volume claim updated to work with Kubernetes 1.6
* Spark driver/executor memory can be set from Luigi and seldon-cli
* Kafka stream processing simplified for API stats on dashboard

* Google Cloud ingress to allow TLS security over external IP to Seldon API on Google Container Platform
* Updated Kafka-stream analytics using 0.10.2.1 of kafka/kafka-streams and running as Kubernetes Deployments
* Several small bug fixes

**Product Naming Update**
Seldon’s current open-source platform has been renamed Seldon Core to allow us to distinguish our open-source machine learning deployment platform from a new product that we are currently working on. We will share more info with you soon and will reach out to this group for beta testers.

Seldon Core 1.4.4. includes various updates:

* [Kubernetes persistent volumes](http://docs.seldon.io/install.htmlstorage) are abstracted by using Persistent Volume Claims. Fixes issue 50.
* All components run as Kubernetes Deployments. Fixes issue 49.
* Initial beta ability to run [external Google MySQL](http://docs.seldon.io/install.htmlmysql) server outside Kubernetes.
* [Updated luigi configuration](http://docs.seldon.io/content-recommendation-guide.htmlmodel) for describing ML pipelines and built-in algorithms.

Thanks to colemickens for submitting both of the above issues!

Bug fix release for kubernetes configuration

Maintenance release.
Docs for anomaly detection demo updated and python libraries in docker fixed.

Maintenance release with updates to [built-in recommendation algorithms](http://docs.seldon.io/content-recommendation-models.html).
Initial code for [Anomaly Detection](https://github.com/SeldonIO/seldon-server/blob/master/python/seldon/anomaly/AnomalyDetection.py).