r3.xlarge has same compute as c3.xlarge, 4x as much ram and almost no demand, so it costs less atm(0.0321 vs 0.0322 :) Judging by spot price history the price on these doesn't fluctuate as much as other nodes. It should get a higher perf constant than m3.xlarge due to more ram.
I'm going to take a look at this. The plan is to test it on some on-demand instances first and then push to spot instances. Any objections against performance constant set to 1.2? Our current values are here: http://hg.mozilla.org/build/cloud-tools/file/dbdf3c0695e8/configs/watch_pending.cfg#l39
Assignee: nobody → rail
Created attachment 8437122 [details] [diff] [review] r3.diff bld-linux64-ec2-001 is r3.xlarge now, will check https://secure.pub.build.mozilla.org/builddata/reports/slave_health/slave.html?name=bld-linux64-ec2-001 We can land this change tomorrow.
Attachment #8437122 - Flags: review?(catlee)
Comment on attachment 8437122 [details] [diff] [review] r3.diff https://hg.mozilla.org/build/cloud-tools/rev/530d17bd5a9c
Attachment #8437122 - Flags: checked-in+
We have ~100 r3.xlarge running in production now. \o/
Status: NEW → RESOLVED
Last Resolved: 4 years ago
Resolution: --- → FIXED
Component: Platform Support → Buildduty
Product: Release Engineering → Infrastructure & Operations
You need to log in before you can comment on or make changes to this bug.