Host metrics

The Metrics & Health page of a source or an AxoRouter host shows the history of the various metrics Axoflow collects about the host.

The charts also show the events of the host, for example, configuration reloads, or alerts that affect the host.

Filter and arrange host metrics

Use the bar above the charts to select which metrics to show, and for which time period.

• Metrics categories: Hide or show the metrics of a category, for example, System. The category of each metric is under the name of its chart. This change isn’t saved: to change the layout of the metrics, use table_chart Customize List.

• calendar_month Time range: Change the time period of the charts. You can use absolute (calendar) time, or relative time (for example, the last 2 days). To move to the previous or the next period of the same length, select the arrows next to the time period (Previous interval and Next interval).

  Axoflow stores all dates in Coordinated Universal Time (UTC), and automatically converts it to the timezone of set in your browser/operating system.

  Note By default, Axoflow stores metrics for 30 days, unless specified otherwise in your support contract. Contact us if you want to increase the data retention time.

• Hide all alerts / Show all alerts: Hide or show the alerts and other events (like configuration reloads) of the host. By default, the charts show these events.

• notifications_active Alerts: Shows the number of active alerts on the host, and opens the list of alerts.

• table_chart Customize List: Change which charts are displayed, and in which order:

  • To change the order of the charts, drag a chart in the Current Charts list.
  • To hide a chart, select Remove.
  • To display a hidden chart, find it in the Available Charts list, and select Add.

  Then select Apply. These changes are persistent and stored in your profile.

  

• settings_backup_restore Restore default chart order: Show the charts that you saved with Customize List, in the saved order. Use it to undo temporary changes, for example, the categories that you hid. The icon is active only when the displayed charts differ from the saved list.

The settings of the filter bar change the URL parameters of the page, so you can bookmark it, or share a specific view by sharing the URL.

Interact with a host metrics chart

In addition to the controls of the top bar, you can interact with the charts in these ways:

• Hover over the info icon on a colored metric card to display the definition, details, and statistics of the metric.

  

• Select a colored card to hide the related metric from the chart. For example, you can hide unneeded sources on the Log input (bytes) and Log input (events) charts.

• Select library_add_check Show all or Hide all above the cards to show or hide every metric of the chart.

• Select stacked_line_chart Show stacked graph / Show unstacked line graph to switch between a stacked graph and an unstacked line graph.

• On charts that show a capacity (for example, Connections and Disk), select the capacity icon (Show capacity or Hide capacity) to show or hide the capacity line.

• Select an event (for example, an alert or configuration reload) to show its details.

  

• To quickly zoom in on a period, click and drag to select the period to display on any of the charts. Every chart will be updated for the selected period. (To return to the previous state, click the Back button of your browser.)

Host metrics reference

For managed hosts, the following metrics are available, in their default order on the page:

• Connections: Active connections of each log source, compared to their configured maximum.

• CPU: Percentage of time the CPU cores spend busy, averaged over all cores and split by mode, within a time window of 5 minutes.

• Disk: Used storage space on each disk device, compared to its total size.

• Dropped packets (total): UDP packets dropped by the OS at each log source, and log events dropped by each destination, per second. The chart shows these metrics:

  • Dropped UDP packets: Number of UDP packets that the OS dropped before processing, per second, averaged within a time window of 5 minutes.
  • Dropped log events: Number of events dropped from event queues within a time window of 5 minutes.

• Log input (bytes): Log data received by each log source, in bytes per second, averaged within a time window of 5 minutes.

• Log input (events): Log messages received by each log source, in events per second, averaged within a time window of 5 minutes.

• Log output (bytes): Log data sent to each destination, in bytes per second, averaged within a time window of 5 minutes.

• Log output (events): Log messages sent to each destination, in events per second, averaged within a time window of 5 minutes.

• Log memory queue (events): Number of messages waiting in the memory queue of each destination.

• Log disk queue (events): Number of messages waiting in the disk queue of each destination.

• Log memory queue (bytes): Memory used by messages waiting in the memory queue of each destination.

• Log disk queue (bytes): Disk space used by the disk queue of each destination, compared to its capacity, plus the memory caching it. The chart shows these metrics:

  • Disk queue bytes: Total bytes of data waiting in each disk queue, and their configured capacity.
  • Disk queue memory cache bytes: Amount of memory used for caching disk-based queues.

• Memory: Memory in use, compared to the total memory reported by the OS (including reclaimable caches and buffers).

• Network input (bytes): Incoming network traffic on each network interface, in bytes per second, averaged within a time window of 5 minutes.

• Network input (packets): Incoming network packets on each network interface, per second, averaged within a time window of 5 minutes.

• Network output (bytes): Outgoing network traffic on each network interface, in bytes per second, averaged within a time window of 5 minutes.

• Network output (packets): Outgoing network packets on each network interface, per second, averaged within a time window of 5 minutes.

• Event delay (seconds): Average time from receiving a message until it’s sent to each destination. The chart is hidden by default. To show it, add it in Customize List.