Setting up Graphite to receive performance data from Unified Manager
PDF of this doc site
- Install Unified Manager on VMware vSphere systems
- Install Unified Manager on Linux systems
- Install Unified Manager on Windows systems
Perform configuration and administrative tasks
- Configuring Active IQ Unified Manager
- Using the maintenance console
Monitor and manage storage
- Monitoring and managing clusters from the dashboard
- Provisioning and managing workloads
Manage events and alerts
- Managing events
Monitor and manage cluster performance
- Navigating performance workflows in the Unified Manager GUI
- Monitoring cluster performance from the Performance Cluster Landing page
- Monitoring performance using the Performance Inventory pages
- Monitoring performance using the Performance Explorer pages
- Analyzing performance events
Monitor and manage cluster health
Common Unified Manager health workflows and tasks
- Monitoring and troubleshooting data availability
- Managing backup and restore operations
- Common Unified Manager health workflows and tasks
Protect and restore data
- Creating and troubleshooting protection relationships
Generate custom reports
- Sample custom reports
Graphite is an open software tool for gathering and graphing performance data from computer systems. Your Graphite server and software must be configured correctly to receive statistical data from Unified Manager.
NetApp does not test or verify specific versions of Graphite or other third-party tools.
After you have installed Graphite according to the installation instructions, you need to make the following changes to support statistical data transfer from Unified Manager:
/opt/graphite/conf/carbon.conffile, the maximum number of files that can be created on the Graphite server per minute must be set to 200 (
MAX_CREATES_PER_MINUTE = 200).
Depending on the number of clusters in your configuration and the statistics objects you have selected to send, there might be thousands of new files that need to be created initially. At 200 files per minute it might take 15 minutes or longer before all metric files are initially created. After all the unique metric files have been created, this parameter is no longer relevant.
If you are running Graphite on a server deployed using an IPv6 address, the value for LINE_RECEIVER_INTERFACE in the
/opt/graphite/conf/carbon.conffile must be changed from “0.0.0.0” to “::” (
LINE_RECEIVER_INTERFACE = ::)
retentionsparameter must be used to set the frequency to 5 minutes and the retention period to the number of days that is relevant for your environment.
The retention period can be as long as what your environment allows, but the frequency value must be set to 5 minutes for at least one retention setting. In the following example, a section is defined for Unified Manager using the
patternparameter, and the values set the initial frequency to 5 minutes and the retention period to 100 days:
pattern = ^netapp-performance\..
retentions = 5m:100d
If the default vendor tag is changed from “netapp-performance” to something different, that change must be reflected in the
patternparameter as well.
If the Graphite server is unavailable when the Unified Manager server is attempting to send performance data, the data is not sent and there will be a gap in collected data.