Epic Goal

• Validate the benefit of isolating etcd in Data Disks on Azure deployments

Why is this important?

• The motivation of this research was the fact that, historically, Azure has had disk performance issues observed by our customers and brought to us through customer cases.  Even with the numerous improvements applied over time and the amount of resources allocated, we still observe issues pointing to very slow (1+s) requests in etcd logs, even though the storage resources allocated for the initial cluster is almost double when compared to the ones for AWS, for example

Scenarios

1. ...

Acceptance Criteria

• TBD
•  
•  
• ...

Dependencies (internal and external)

1. ...

Previous Work (Optional):

1. …

Open questions::

1. …

Done Checklist

• TBD