Thresholds

Define threshold limits for node resource utilization and trigger events
Tags:

  3 minute read  

Thresholds provide a way to trigger events when different measurements exceed a given value. Thresholds configured at the node level can override Domain Thresholds.

Example Thresholds page on a node

Node Level Threshold Status

Field NameDescription
Status
  • Disabled (default) - Only thresholds defined at the domain level will be used for generating events.
  • Enabled - Thresholds defined at the node level will be used for generating events, including overriding any domain-level values.

Configuring Node Level Thresholds

To configure thresholds at the node level:

  1. Navigate to the node in the portal.
  2. In the left-side navigation bar find Thresholds in the System section and click it.
    System > Thresholds
  3. Change the Status to “Enabled”.
  4. Click the + button on the far right to add either a Load or Network threshold.
  5. Provide the required fields.
  6. Optionally, repeat with any other thresholds you wish to set.
  7. Click save.

Load Thresholds

Load thresholds measure the health of the node itself.

Load Metrics

Metric TypeDescriptionDefault Value
CPU UsageMonitors percent of CPU usage across all cores95% for 10 minutes
Memory UsageMonitors percent of total memory (RAM) used90% for 30 minutes
Disk UsageMonitors percent of total disk usage for the root partition80% for 1 minute
Embrionic FlowsMonitors the number of TCP flows (connections) that are in the embryonic state (waiting for ACK)none
JVM HeapMonitors the percent of allocated JVM memory usednone

Load Fields

Field NameDescription
NameThe name of the threshold. This will be available in generated events.
TelemetryThe metric to monitor. Options are CPU usage (%), memory usage (%), disk usage (%), and embryonic flows (absolute count).
ThresholdThe value that must be exceeded for an event to be generated.
DurationThe time period to measure. If the threshold is exceeded for this duration, an event will be generated.

Network Thresholds

Network thresholds measure the health of the network from the node’s perspective.

Network Metrics

Metric TypeDescription
Latency (ms)Monitors the round trip tunnel latency between this node and the target
Bandwidth IN Usage (Mbps)Monitor the amount of received bandwidth on the specified interface
Bandwidth Out Usage (Mbps)Monitors the amount of sent bandwidth on the specified interfaces

Network Fields

Field NameDescription
NameThe name of the threshold. This will be available in generated events.
TelemetryThe metric to monitor. Currently only latency (measured in milliseconds) is available.
ThresholdThe value that must be exceeded for an event to be generated.
DurationThe time period to measure. If the threshold is exceeded for this duration, an event will be generated.
Target
  • For Latency - The target node to measure the latency to. Each node will measure the latency to the target node.
  • For Bandwidth In/Out - The network interface to monitor usage on.