Set up the conditions where a cluster sends an alert.
Edit the configuration file
From the DataIQ browser, select
Settings > Storage monitoring settings > Configure metric threshold alerts.
Configuration file conventions:
● Use spaces for indentations, not tabs.
● Ensure indentation matches exactly.
● When using boolean values and numbers, do not use quote marks. For all other values, use quote marks.
Configuration file settings
Table 1. protocol_latency: cluster
cluster: <clustername>
or
cluster: [<clustername_1>, <clustername_2>...]
Optional: Select clusters by name. To choose multiple clusters, separate the cluster names with commas and a space. May be further refined by protocol, condition.
Example:
protocol_latency:
- cluster: [cluster-1-name, cluster-2-smartconnect-ip]
protocol: any
condition:
above: 1000
To select any cluster, either do not include this setting, or set this value to
any. Example:
protocol_latency:
- cluster: any
protocol: any
condition:
above: 1000
Required for each instance of
cluster. Select clusters by protocol. To choose multiple protocols, separate the protocols with commas and a space. Example:
protocol_latency:
- cluster: any
protocol: [nfs3, nfs4, smb1, smb2]
To select any protocol, use
any. Example:
protocol_latency:
- cluster: any
protocol: any
condition:
above: <latency>
Required for each instance of
cluster. Sets maximum allowed cluster latency before generating an error, in milliseconds. Example:
protocol_latency:
- cluster: any
protocol: any
condition:
above: 1000
condition and
above are required and cannot be empty.
Optional: Select individual clusters by the cluster's primary IP address. To choose multiple clusters, separate the IP addresses with commas and a space. Example:
protocol_latency:
- cluster: any
client: [192.168.0.1, 192.168.0.2]
protocol: any
condition:
above: 1000
To select any client, do not include this value.
Node-level reporting is not supported.
When cluster is added in DataIQ using the first node DNS/IP, it is also required to follow the same record in alert rules.
client: any is not supported.
Basic example
Send an alert when any protocol latency on any cluster is greater than 1000 milliseconds.
protocol_latency:
- cluster: any
protocol: any
condition:
above: 1000
Example: Different protocols
It is recommended to specify different threshold for different protocols based on the actual needs.
protocol_latency:
- cluster: any
protocol: [nfs3, nfs4]
condition:
above: 1000
- cluster: any
protocol: [smb1, smb2]
condition:
above: 2000
Example: Cluster list and protocol list
Instead of cross joining all clusters with all protocols, it is also recommended to specify both the cluster list and protocol list per actual needs.