Status: In review
This SIG aims to improve the experience of developing, operating, and using auto-scaling and its related features (like metering, cluster schedule, life cycle management), and to coordinate efforts across projects and communities (like k8s cluster auto-scaling on OpenStack). The SIG also provides a central place to put tests, documentations, and even common libraries for auto-scaling features.
The SIG is expected to focus more on auto-scaling user workloads; however work on auto-scaling infrastructure is also welcome, especially considering that user workloads in an undercloud are actually infrastructure in the corresponding overcloud.
OpenStack provides multiple methods to auto-scale your cluster (Like using Heat AutoScalingGroup, Senlin Cluster, etc.). However without general coordination across projects, it may not be easy for users and ops to achieve auto-scaling on OpenStack. Developers tend to be focused on individual projects rather than cross-project integration. Most of the components required by auto-scaling already exist within OpenStack, but we need to provide a more simple way for users and ops to adopt auto-scaling. And allowing developers to coordinate together instead of implement something all over again.
- Set up SIG, init doc, meeting, ml
- Create Forum and Auto-scaling SIG PTG sessions
We're currently working on a structure to make everyone more easier to getting involved. Our initial git repository, documentation, and other information will be up soon.
Community Infrastructure / Resources
- Wiki: this page
- SIG StoryBoard (for an authoritative list of all ongoing work within the SIG)
- openstack-discuss mailing list; use the
- IRC channel: #openstack-auto-scaling on Freenode IRC
- Regular IRC meetings have not been arranged yet
- patch reviews (gerrit)
- Rico Lin
- Duc Truong
- Zane Bitter
We welcome anyone who willing to join and help. And we surely are willing to re-elect chairs as soon as we got our init setup done.
- Events at Berlin Forum, November 2018
This is a good idea, but we need some more discussion on who and how we can bridge this system up.