Current Quota

808 Nodes

• 8GM RAM
• 8vCPU at 2.4GHz (more of less depending on speed)
• Public IP Address
• 80GB of storage

Providers

• Rackspace: 60(DFW), 195(ORD), 45(IAD)
• OVH: 79(GRA1), 159(BHS1)
• IBM Bluebox: 39
• Internap: 56
• OpenStack Innovation Center: 100
• Vexxhost: 25

Also

• HPE Regions of Infra Cloud: 600? (100 compute nodes)

Multi-cloud FTW

• Earier this year, HPE Sunset their public cloud
• Over the last three weeks, we lost RAX ORD, OVH BHS and vexxhost
• Last week we had Floating IP leak problems with OSIC and Bluebox

... and we still limped along.

Nodes in Use

808 Node Max Quota

Averaged over the entire time period in question

• 1 year average : 381 nodes
• 6 month average: 403 nodes

Time spent on Projects

Time spent running jobs for projects of groups of projects.

Does not account for multi-node

'Integrated' is the set of projects that were in the integrated release back when we had one

Feb 28 -- Mar 9

Milestone 3, feature/requirements freeze

Total time all projects:        226042021
Total time official projects:   224206325  99.2%
Total time unofficial projects:   1835696   0.8%
Total time integrated projects: 127325612  56.3%
Total time non-voting jobs:      76931278  34.0%
      

Mar 11 -- Mar 21

RC1 target

Total time all projects:        131665133  (58% of usage of previous period)
Total time official projects:   129752051  98.5%
Total time unofficial projects:   1913082   1.5%
Total time integrated projects:  59619553  45.3%
Total time non-voting jobs:      46111506  35.0%
      

Mar 28 -- April 7

Final RC, release

Total time all projects:        186115455  (141% of usage of previous period)
Total time official projects:   182577152  98.1%
Total time unofficial projects:   3538303   1.9%
Total time integrated projects:  91231261  49.0%
Total time non-voting jobs:      60635432  32.5%
      

Control Plane

Quick shout-out to Rackspace, who provide our control plane all by themselves

• 110 Servers
• 30.6 TB of Block Storage

What is needed?

• More public cloud resources are always welcome
• Once Infra Cloud is back online, having discussions about more hardware contributions is worth having
• zuul v3 this cycle will make multi-node MUCH more powerful and flexible, so demand is likely to increase.
• zuul v3 this cycle will also enable consuming non-vm resources

DCO implementation status

• Getting off of Launchpad OpenID is the main todo list item
• Bug Tracking is the sticky widget
• Design Summit session on Phabricator and Storyboard on Wednesday
• "Why don't you just ..."