CycleCloud version 7.8.0

This release provides improved user management, updates Slurm autoscaling, and increases the orchestration and management efficiency.

New Features:

CycleCloud can now manage multiple user accounts on each cluster node. Users are dynamically added/removed on nodes based on cluster permissions
User's SSH public keys are now validated when adding them to the user profiles in the UI
Automatic node user management may be enabled/disabled via a global setting
Cluster nodes can be pinned to a specific availability zone within a region
The autoscaling API now supports a simple core-based allocation
The autoscaling API supports starting/terminating nodes by hostname
Added a Python API for easy interaction with the CycleCloud REST API
Estimated VM costs now include disk pricing as well as compute pricing
The cyclecloud connect command takes a hostname to connect to
The UI and CLI now displays the nodes hostnames
New node attributes NameFormat and NameOffset allow customized array node names
Nodes with StartAutomatically = false will stay in the 'off' state when they are added.
Clusters of up to 5000 nodes are now supported

Some transient authentication errors were showing as failures in the UI
Public IP addresses are properly cleared in the UI when AssociatePublicIpAddress is set to false on a node
Deallocated nodes no longer show as full price
Hb, Hc and H16 Promo SKUs now select the CentOS HPC image correctly
In some cases, the cluster would show "Cell execute-XYZ not found" errors
SSH proxy failed when connecting through bastion hosts
Detection of selinux policy failed on Enterprise Linux version 6
In rare cases, form fields would not be saved properly from the UI
Fixed an issue with interrupted threads sometimes preventing records from being saved to the datastore
Authentication error detail was being omitted from the UI and logs
The "Add Nodes" button was not shown properly to non-administrator users
Cluster parameter values containing "$" were not being evaluated correctly