Description of requirements: New Metering Calculation for Middleware Products running on OpenShift
Filing / continuation of discussions from before shutdown with Oleg and Greg. PRIMARY METERING ALGORITHM Per universal clock our, report the maximal value for instantaneous CPU core usage across containers bearing a given label inside OpenShift. Visual Aid - https://docs.google.com/presentation/d/13O1vQW_gJR_S4L16C9T2XOnz0xp0gzJInR8EkTo3JBQ/edit#slide=id.g27cdf96659_2_480 SIGNIFICANT REPORTING CONCERNS Per label the report should be capable of grouping by project, sorting usage of common images into "buckets." Customers will use these groupings to index use against cost centers requesting and operating the projects. Assignation of a literal rate to this usage, integrating with chargeback features, is considered a nicety - but not essential for subscription reporting needs. ADDITIONAL REPORTING CONCERNS The middleware business unit is in discussions with various parties to use memory as the active metric for calculating subscription "size." The ability to map the "hourly instantaneous maximum" calculation against memory usage would be an extremely wise feature to build in. Further, the middleware business unit is contemplating the use of allocation rather than literal usage for calculating subscription "size." The ability to map the "hourly instantaneous maximum" calculation against allocation of both memory AND core usage would be an excellent addition, if possible. DATA RETENTION The middleware business unit is seeking feedback from engineering regarding what is possible for data retention. Primary use of retained data would be explaining or justifying an hourly value/calculation to a customer. Ideal data retention would preserve all sampled data for a relatively short period of time, while holding a condensed "rollup" for a longer period. The rollup can be as simple as "For the 1400-1500 hour on January 23, 2018 the calculated maximum was 24 cores at 1437. Pod 1 from image X = 2.4 cores. Pod 2 from image X = 2.6 cores. Pod 1 from image Y = 7 cores. Pod 2 from image Y = 7.4 cores. Pod 3 from image Y = 2.1 cores. Pod 1 from image Z = 2.2 cores."
Dan, do you have a live environment that we can use to construct and verify the report are we work on it? This will help us tremendously so the we can focus on the building the report and not lose time trying to build an env and collect data.
Hopefully this will do for you: https://docs.google.com/document/d/1US6GiBsRqJ4syJGL_cCQAZr3pzu8U83eaecwlVznYpU/edit Doc should be viewable to you. Ansible QE stood this up for us - credit where due. :)
https://github.com/ManageIQ/manageiq/pull/16826
https://github.com/ManageIQ/manageiq/pull/16852
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/4ae0dfe26f93c7947158dfbdad9a0352878a746b commit 4ae0dfe26f93c7947158dfbdad9a0352878a746b Author: Gregg Tanzillo <gtanzill> AuthorDate: Fri Jan 19 06:48:31 2018 -0500 Commit: Gregg Tanzillo <gtanzill> CommitDate: Fri Jan 19 06:48:31 2018 -0500 Move `sqlite3` into main section of Gemfile Need Sqlite for some new tools currently under development https://bugzilla.redhat.com/show_bug.cgi?id=1531561 Gemfile | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/8969f7b27d1b471c2bd58bfb1abbc6fba1ee2308 commit 8969f7b27d1b471c2bd58bfb1abbc6fba1ee2308 Author: Gregg Tanzillo <gtanzill> AuthorDate: Tue Jan 16 07:01:54 2018 -0500 Commit: lpichler <lpichler> CommitDate: Wed Jan 17 14:29:55 2018 +0100 Version 1 capture script for radar project Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1531561 tools/capture_radar.rb | 39 +++++++++++++++++++++++++++++++++++++++ 1 file changed, 39 insertions(+) create mode 100644 tools/capture_radar.rb
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/fd6bce26bff4713a472980b367a46a26fcbd37b2 commit fd6bce26bff4713a472980b367a46a26fcbd37b2 Author: Gregg Tanzillo <gtanzill> AuthorDate: Tue Jan 16 08:07:26 2018 -0500 Commit: lpichler <lpichler> CommitDate: Wed Jan 17 14:29:55 2018 +0100 Added container project Metric rollups are separated by timestamp, CustomAttribute (label) and project name (resource_name) Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1531561 tools/capture_radar.rb | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-)
New commit detected on ManageIQ/manageiq/master: https://github.com/ManageIQ/manageiq/commit/676dc0160dfee24db1a7244f06db24dac49b8982 commit 676dc0160dfee24db1a7244f06db24dac49b8982 Author: Gregg Tanzillo <gtanzill> AuthorDate: Tue Jan 16 07:41:18 2018 -0500 Commit: lpichler <lpichler> CommitDate: Wed Jan 17 14:29:55 2018 +0100 Persist max CPU cores used by label as MetricRollup Fixes https://bugzilla.redhat.com/show_bug.cgi?id=1531561 tools/capture_radar.rb | 20 ++++++++++++++------ 1 file changed, 14 insertions(+), 6 deletions(-)