Breadcrumbs

System Health

Overview

The System Health screen enables the user to monitor the health and resource usage of the devices deployed in the selected Site. It shows key system metrics of the selected device, such as CPU, memory, disk and swap utilization, temperature and network flow count, in the form of charts and tables.

This screen helps the user to identify issues such as high resource usage, overheating or abnormal network activity on a device before they affect the network performance of the Site.

In the System Health screen, the user can:

  • Select the device for which the system health metrics are to be viewed.

  • Select the time range and the interval of the data shown in the charts.

  • View the system health charts of the selected device.

  • View the Flow Count Table of the clients connected to the device.

Access the System Health Screen

  1. Select Sites from the left menu bar of Kognitive Cloud.

  2. In the Organization field, select the required Organization or sub-organization from the dropdown menu.

  3. In the Site field, select the required Site from the dropdown menu.

  4. Click the System Health tab to open the System Health screen.

Select the Device

  • Click the select device field to open the dropdown menu. The dropdown lists all the devices deployed in the selected Site.

  • Select the required device.

The charts and the Flow Count Table are populated with the data of the selected device. If no data is available, No Data is shown in the charts.

Select the Time Range and Interval

  • Click the time range dropdown at the top right corner of the screen and select the required time range (Last 15 min, Last 1 Hr, Last 6 Hr, Last 12 Hr, Last 1 Day, Last 7 Days, Last 30 Days, MTD or a custom From and To range).

  • Click the interval dropdown above the charts and select the required interval - 5m, 10m, 20m, 1h, 6h, 12h, 1d or 7d. The interval determines the granularity of the data points plotted in the charts.

  • After zooming into a chart, click Reset Zoom to return the charts to their original view.

System Health Charts

The System Health screen shows seven charts for the selected device. All the charts share the same layout:

  • The horizontal axis shows the time instances within the selected time range (for example, 21 Sep to 19 Oct when Last 30 Days is selected).

  • The vertical axis shows the value of the metric in its unit (Percentage (%), Counts or Temperature (°C)).

  • Each dot on the line is one data point. The number of data points depends on the selected interval. For example, when the time range is Last 30 Days and the interval is 12h, one data point is plotted for every 12 hours, which gives about 60 data points per chart. Select a smaller interval to see more detail, or a larger interval to see the overall trend.

  • The legend at the top right corner of each chart names the data series plotted in the chart.

  • All the charts use the same time range, so a change in one chart can be compared with the other charts at the same time instance. For example, a drop in Swap Utilization at the same time as a spike in Number of Sockets and Io Rx Drops can point to a single event on the device, such as a restart.

CPU Utilization

The CPU Utilization chart shows how much of the processing capacity of the device is being used over time. The CPU handles all the work done by the device, such as routing traffic, applying traffic policies and firewall rules, and running the device services.

  • Vertical axis - Percentage (%) of CPU capacity in use, from 0 to 100.

  • Legend - CPU - Overall CPU utilization of the device.

How to read the chart:

  • Each data point shows the CPU utilization of the device at that time instance.

  • A low and steady line (for example, a value that stays around 20%) indicates that the device has sufficient processing capacity for the current load.

What to look for:

  • Sustained high values (consistently close to 100%) indicate that the device is overloaded. This can slow down traffic processing and increase latency for the users on the site.

  • Sudden spikes can be caused by a short burst of traffic or a temporary process on the device. Compare the time of the spike with the Number of Sockets chart and the Events screen to find the cause.

Memory Utilization

The Memory Utilization chart shows how much of the memory (RAM) of the device is being used over time. The device uses memory to run its services and to keep track of the active network connections.

  • Vertical axis - Percentage (%) of memory in use, from 0 to 100.

  • Legend - Memory - Overall memory utilization of the device.

How to read the chart:

  • Each data point shows the memory used by the device at that time instance.

  • It is normal for a device to use a stable portion of its memory (for example, a value that stays around 60%) while it is in operation.

What to look for:

  • A line that rises continuously over time without coming down can indicate that a service on the device is not releasing memory.

  • Sustained values close to 100% indicate that the device is running out of memory. When memory is full, the device uses swap space instead, which is shown in the Swap Utilization chart.

Disk Utilization

The Disk Utilization chart shows how much of the storage space of each partition (mount point) on the device is being used over time. Each partition stores a different type of data, so each one is plotted as a separate line.

  • Vertical axis - Percentage (%) of the partition space in use, from 0 to 100.

  • Legend - One line for each partition of the device. The partitions are:

  • boot - Partition that stores the files required to start the device.

  • boot/efi - Partition that stores the EFI boot files of the device.

  • data - Partition that stores the application data of the device.

  • etc/k4data - Partition that stores the Kognitive configuration data of the device.

  • media/root/ro - Read-only root partition of the device.

  • media/root/rw - Read-write root partition of the device.

  • var/log - Partition that stores the log files of the device.

How to read the chart:

  • Each line shows the percentage of space used in that partition at each time instance. The lines usually stay flat, because the used space of most partitions changes slowly.

What to look for:

  • A line that rises steadily, especially for var/log or data, indicates that the partition is filling up.

  • A partition close to 100% can stop the device from writing logs or saving data, and can affect the services running on the device. Contact Kognitive Networks support if a partition is close to full.

Number of Sockets

The Number of Sockets chart shows the number of network connections that the device is tracking over time. Every connection that passes through the device, for example, a client browsing a website or an application connecting to a server, is tracked by the device as an entry in its connection tracking table.

  • Vertical axis - Counts - number of tracked connections.

  • Legend - Contrack - Number of entries in the connection tracking (conntrack) table of the device.

How to read the chart:

  • Each data point shows the number of connections tracked by the device at that time instance.

  • The value depends on the number of clients on the site and how they use the network, so it normally moves up and down within a range (for example, between 1100 and 1350).

What to look for:

  • A sudden spike (for example, a jump from about 1200 to about 1800) indicates a sudden increase in the number of connections. This can be caused by a client or an application that opens a large number of connections. Use the Flow Count Table to find the clients with the highest flow count.

  • Consistently high values can affect the performance of the device and increase CPU and memory utilization.

Io Rx Drops

The Io Rx Drops chart shows the number of incoming (received) packets that were dropped by the device over time. A packet is dropped when the device cannot accept or process it, for example, when the receive buffer of the interface is full.

  • Vertical axis - Counts - number of dropped incoming packets.

  • Legend - Io Rx Drops - Number of received packets dropped by the device.

How to read the chart:

  • Each data point shows the number of received packets dropped at that time instance.

  • In normal operation, the value stays at or close to 0.

What to look for:

  • Any value above 0 indicates that the device dropped incoming packets during that period. A short, isolated increase (for example, a value of 1 for one or two data points) is usually not a concern.

  • Frequent or continuous drops can cause slow connections, retransmissions and poor user experience. Compare the time of the drops with the CPU Utilization and Number of Sockets charts to check whether the device was overloaded at that time.

Swap Utilization

The Swap Utilization chart shows how much of the swap space of the device is being used over time. Swap space is a part of the disk that the device uses as extra memory when the memory (RAM) is full or when data in memory is not being actively used.

  • Vertical axis - Percentage (%) of swap space in use, from 0 to 100.

  • Legend - Swap - Overall swap utilization of the device.

How to read the chart:

  • Each data point shows the swap space used by the device at that time instance.

  • Swap usage usually increases slowly while the device is running and resets when the device restarts. For example, a sudden drop from about 40% to about 5%, followed by a gradual increase, usually indicates that the device was restarted at that time.

What to look for:

  • Swap space is slower than memory, so high swap utilization can slow down the device. Read this chart together with the Memory Utilization chart: high memory utilization together with high swap utilization indicates that the device needs more memory than is available.

  • Sustained values close to 100% indicate that the device is running out of both memory and swap space.

Temperature

The Temperature chart shows the operating temperature of the device over time. Devices installed on vessels and other remote sites can be exposed to heat, poor ventilation or direct sunlight, which can affect their performance and lifetime.

  • Vertical axis - Temperature (°C).

  • Legend - Temperature - Operating temperature of the device.

How to read the chart:

  • Each data point shows the temperature of the device at that time instance.

  • A steady line (for example, a value that stays between 35°C and 38°C) indicates that the device is operating in a stable environment.

What to look for:

  • A temperature that rises steadily or stays high can indicate poor ventilation, a blocked air flow or high ambient temperature at the installation location. Check that the device is installed according to the installation guidelines.

  • Overheating can cause the device to slow down or restart. Compare the time of a temperature increase with the CPU Utilization chart, because high CPU load also increases the temperature of the device.

Zoom into the Charts

  • Zoom into a chart to view the data of a shorter period in more detail.

  • Click Reset Zoom at the top right corner of the screen to return the charts to their original view.

Flow Count Table

The Flow Count Table shows the number of network flows (connections) created by each client connected to the selected device. This helps the user to identify the clients that create an unusually high number of connections.

The table shows the following information:

  • IP Address - IP address of the client

  • MAC Address - MAC address of the client

  • Average Flow Count - Average number of flows created by the client during the selected time range

  • Max Flow Count - Maximum number of flows created by the client during the selected time range

  • Enter the required text in the Type something search field to search the table.

  • Click a column header to sort the table based on that column.

  • Click the Download icon next to the search field to download the Flow Count Table.

  • Use the Items per page dropdown to select the number of rows shown on each page, and use the page dropdown or the arrows to move between the pages.