search

redhat-performance / quads

:calendar: The infrastructure deployment time machine
https://quads.dev
GNU General Public License v3.0
87 stars 36 forks source link

Populate per-port bandwidth graphing during assignments #85

Closed sadsfae closed 1 week ago

sadsfae commented 7 years ago

Currently we have a stub in the wiki generation for each host called graph. This was left in place to late come through and add collectd-snmp port utilization and bandwidth metrics which would point to Grafana graphs for each host.

We would need to have dashboards automatically generated for all interfaces for each unique cloud assignment and owner. Retention can be set on the Graphite side.

screenshot_2017-05-12_15-44-12

bengland2 commented 7 years ago

what about having a link to a graph of bandwidth on uplinks for each rack entry here next to each rack's name (e.g. "Rack 08"). This would be simpler to measure and would be useful for troubleshooting "noisy neighbor" network problems - since we can easily determine who we are sharing a rack with. Joe's ideas for measuring network throughput of a cloud is a good one but I suspect it will be a while before we get to it, and knowing the cloud throughput doesn't necessarily mean you know whether a particular rack is overloading or not.

sadsfae commented 12 months ago

Related-to: https://github.com/redhat-performance/quads/issues/228

sadsfae commented 3 months ago

We are now populating per-cloud bandwidth and power usage within Grafana (outside of QUADS). I will leave this open and move it to 2.0 roadmap for implementing an optional place for wiki generation to include it for people that have similar setups.

sadsfae commented 1 week ago

Closing this issue, it's quite old and it also arguably extends farther than we can maintain into third-party platform support.