Monitoring System Performance

Using the System > Performance Page

The System > Performance page displays basic AirWave hardware information as well as resource usage over time. AirWave logs performance statistics such as load average, memory and swap data every minute.

The historical logging is useful to determine the best usable polling period and track the health of AirWave over time.

The page is divided into the following sections:

System Information
Performance Graphs
AMON Statistics
Redis Statistics
Database Statistics
Disk Space

Figure 1 illustrates this page, and Table 1 describes fields and information displayed.

Figure 1 System > Performance Page Illustration (Partial Screen)

Table 1: System > Performance Page Fields and Graphs
Field	Description
System Information
Current Time	Displays the current time on the AirWave server.
CPU(s)	Basic CPU information as reported by the operating system.
Memory	The amount of physical RAM and Swap space seen by the operating system. Refer to the HPE Aruba Networking Management Software (AirWave) Server Sizing Guide for hardware requirements.
Kernel	The version of the Linux kernel running on the box.
Device Polling	Displays some AP/Device polling statistics.
Performance Graphs
System Load Average	The number of jobs currently waiting to be processed. Load is a rough metric that will tell you how busy a server is. A typical AirWave load is around 2-3 times the number of CPU cores you have in your system. A constant load of 4x to 5x is cause for concern. A load above 6x is a serious issue and will probably result in AirWave becoming unusable. To lower the load average, try increasing a few polling periods in the Groups > Basic page.
System Memory Usage	The amount of RAM that is currently used broken down by usage. It is normal for AirWave to have very little free RAM. Linux automatically allocates all free RAM as cache and buffer. If the kernel needs additional RAM for process it will dynamically take it from the cache and buffer.
System Swap Usage	The amount of Swap memory used by AirWave. Swap is used when there is no more free physical RAM. A large performance penalty is paid when swap is used. If your AirWave consistently uses swap, you should consider installing additional RAM.
System Disk Throughput	The rate of reading and writing from and to the disk in bytes per second.
System Disk IOPs	The number of disk reads and writes per second.
System Disk Outstanding I/O Requests	The average number of outstanding I/O requests (queue depth). If it's high, it means that I/O requests (disk reads/writes) aren't being serviced as fast as they're being asked for.
System Disk Utilization	The amount of data read from the disk and written to the disk.
System CPU Utilization	The percentage of CPU that has been used by the user and the system as well as the amount that was idle.
Process Counts by Service	This breaks down network usage based on Web server, database, AirWave Service, and VisualRF processes.
Average Delay Time by Queue Type	This shows the queue time for Async logger clients and RAPIDS processing.
I/O Throughput by Worker/by Service	Displays reads and writes for workers (AirWave services, database, VisualRF, web server, RRD tool and AWRRD tool) and for services (AirWave, VisualRF and web server).
CPU Utilization by Worker/by Service	Displays reads and writes for workers (AirWave services, database, VisualRF, web server, RRD tool and AWRRD tool) and for services (AirWave, VisualRF and web server).
System Network Usage	All traffic in and out measured in bits per second of your primary network interface (Eth0 being the most common).
Usage by Protocol	Displays the amount of traffic used by Telnet, HTTPS and SNMP used by your primary network interface (Eth0 being the most common).
Syslog	Displays the incoming Syslog packets on the AMP.
SNMP Traps	Displays the number of SNMP Trap packets in your network over the last two hours, day, week, month, and year
Legacy SNMP Fetcher Requests	The number of SNMP get and walk requests per second performed by the legacy (v1 and v3) SNMP fetcher.
Legacy SNMP Fetcher Responses	The number of SNMP OIDs received per second performed by the legacy (v1 and v3) SNMP fetcher.
High Performance SNMP Fetcher Requests	The number of SNMP get and walk requests per second performed by the high performance SNMP (v2c) fetcher.
High Performance SNMP Fetcher Responses	The number of SNMP OIDs received per second performed by the high performance SNMP (v2c) fetcher.
Redis Statistics
Redis Activity	Use this chart under the supervision of HPE Aruba Networking support to troubleshoot Redis activity. Click any point in the chart to view Redis activity over the past day, week, month or year.
Redis Used Memory	Use this chart under the supervision of HPE Aruba Networking support to troubleshoot Redis memory issues. Click any point in the chart to view the total number of bytes used by the Redis process over the past day, week, month or year.
Redis Keyspace	Use this chart under the supervision of HPE Aruba Networking support to troubleshoot Redis keys. Click any point in the chart to view Redis Key usage over the past day, week, month or year.
Database Statistics
Top 5 Tables (by row count)	The five largest tables in AirWave. Degraded performance has been noticed for in some cases for tables over 200,000 rows. Decreasing the length of time client data is stored on the AirWave page is recommended if a user/client table exceeds 250,000 rows.
Database Table Scans	The number of database table scans performed by the database.
Database Row Activity	The number of insertions, deletions and updates performed to the database.
Database Transaction Activity	The number of commits and rollbacks performed by the database.
Disk Space
Disk Space	Pie charts that display the amount of used and free hard drive space for each partition. If a drive reaches over 80% full, you may want to lower the Historical Data Retention settings on the AMP Setup > General page or consider additional drive space.

There are several initial steps that you can take to troubleshoot AirWave performance problems, including slow page loads and timeout errors. Initial troubleshooting steps would include the following:

Increasing the polling period settings on the Groups > Basic page.
Increasing the polling period time for groups with routers and switches.
Adding additional memory to the server. Please consult the sizing information in the latest edition of the HPE Aruba Networking Management Software (AirWave) Server Sizing Guide or contact HPE Aruba Networking support for the latest recommendations.