Collectd Gnocchi Status

Overview

This plugin provides valuable insight into your metrics and measures backlog of Gnocchi when installed in an OpenStack environment. It costs an API call to Gnocchi status API every $interval (set in the collectd config file per the plugin). Gnocchi is a component of the OpenStack Telemetry project for Metric-as-a-Service.

Sample Graph

Install / Configuration

1. Assuming you have collectd installed already, append the following plugin details to your collectd.conf config file

   ```
   <LoadPlugin python>
     Globals true
   </LoadPlugin>
   
   <Plugin python>
     LogTraces true
     Interactive false
     Import "collectd_gnocchi_status"
     <Module collectd_gnocchi_status>
       interval 30
     </Module>
   </Plugin>
   ```
   

2. Setup your environment variables in the collectd systemd unit file

   ```
   # Populate the following variables with your stackrc/overcloudrc or openstackrc file
   # Tenant v1&v2 api, project for v3 api
   Environment=OS_TENANT_NAME=admin
   Environment=OS_PROJECT_NAME=admin
   
   Environment=OS_PASSWORD=xxxxxxxxxxxxxxxxxxxxxxxxx
   Environment=OS_USERNAME=admin
   Environment=OS_AUTH_URL=http://x.x.x.x:5000/v2.0
   Environment=OS_CLOUDNAME=overcloud
   Environment=OS_NO_CACHE=True
   # End Environment variables to configure
   ```
   

3. Install plugin

   ```
   [root@overcloud-controller-0 ~]# pip install collectd-gnocchi-status
   ```
   

4. Reload Systemd units

   [root@overcloud-controller-0 ~]# systemctl daemon-reload
   

5. Restart collectd

   [root@overcloud-controller-0 ~]# systemctl restart collectd
   

6. View metrics on Gnocchi in your TSDB

Graphite storage-aggregation.conf Example

Included in the repo is an example Graphite storage-aggregation.conf (sample-storage-aggregation.conf) to ensure that your Gnocchi status metrics are not lost due to your retention sizing (Ex. 10s:7d,60s:90d,1h:180d) < the interval (30s) this plugin is set to collect and publish samples. Simply copy and paste the example, into your /etc/carbon/storage-aggregation.conf file on your Graphite server prior to pushing these metrics and restart carbon-cache to reload this configuration. If the metrics have already been pushed, you must either use whisper-resize (included with Graphite) or remove the corresponding whisper files after restarting carbon-cache to allow carbon to rebuild them with the new parameters in storage-aggregation.conf.

Resources

1. Gnocchi.xyz
2. Collectd.org