eG Monitoring
 

Measures reported by A10ServerTest

Physical servers a.k.a Real servers are those that are bound to a virtual server in a server farm of the A10 Application Delivery Controller. Whenever a client request is received, the virtual server bound to the real server responds to those requests by channelizing the requests to the real servers that are currently available. Since multiple VIPs can be pointed to the same set of real servers, having a good number of supported VIPs presents more flexibility in the architecture and design of the site or application. There may be upto 100 real servers connected to a single virtual IP and the same set of real servers can be pointed to multiple Virtual IPs to provide more flexibility in the architecture and sdesign of the A10 Application Delivery Controller. The A10 Application Delivery Controller installed in large environments often receives thousands of client requests per second, which should be responded without any time delay. In such cases, the virtual IP sends the requests continuously to the available real servers bound to it. If the real server is experiencing any technical glitch or a slowdown or if the real server is currently overloaded, the A10 Application Delivery Controller may not be effective in responding to the client requests thus causing inconsistencies in the load balancing functionality. To avoid such inconsistencies, it is necessary to monitor the health and the request processing details of the real servers. This is where the A10ServerTest exactly helps!

For each real server configured on the A10 Application Delivery Controller, this test continuously monitors the health of the real servers and reveals how well each server processes client requests. In addition, this test detects inconsistencies in load-balancing early on and warns administrators of possible deviations proactively.

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
Health_status Indicates health of this real server.   The values of this measure and their corresponding numeric values are listed below:

Measure Value Numeric Value
Disabled 0
Up 1
Down 2

Note:

By default, this measure reports one of the Measure Values listed in the table above to indicate status of this real server. In the graph of this measure however, the real server status will be represented using the numeric equivalents.
Data_transmitted Indicates the rate at which data was transmitted from this real server during the last measurement period. MB/Sec Compare the values of these measures across the node that is handling maximum traffic.
Data_received Indicates the rate at which data was received by this real server during the last measurement period. MB/Sec
Packets_transmitted Indicates the rate at which the packets were transmitted from this real server during the last measurement period. Packets/Sec Compare the value of these measures across the real servers to identify the real server that is experiencing the maximum traffic.
Packets_received Indicates the rate at which packets were received by this real server during the last measurement period. Packets/Sec
Curr_connections Indicates the number of connections that are currently active on this real server. Number This measure is a good indicator of the load on the real server.
Tot_connections Indicates the total number of connections established on this real server since the start of the A10 Application Delivery Controller. Number  
Connection_rate Indicates the rate at which the connections were established on this real server during the last measurement period. Conns/Sec A sudden increase in the value of this measure indicates an increase in the load on the real server.
Conn_usage Indicates the percentage of connections used by this real server. Percent A value close to 100% indicates that this real server is currently overloaded.
Persistent_conns Indicates the number of connections that were persistent on this real server. Number TCP connections that are kept open after transactions complete are called persistent connections. Persistent connections stay open across transactions, until either the client or the server decides to close them. These connections when reused can significantly reduce the overload on the new connections to the real server.
Peak_conns Indicates the maximum number of connections that were established on this real server since the start of the A10 Application Delivery Controller. Number  
Curr_L7_requests Indicates the number of L7 requests currently processed by this real server. Number Both these measures serve as effective pointers to the L7 requests processing in the A10 Application Delivery Controller.

Layer-7 load balancing, also known as application-level load balancing, is to parse L7 requests in application layer and distribute L7 requests to the servers based on different types of request content, so that it can provide quality of service requirements for different types of content and improve overall performance.
Tot_L7_requests Indicates the rate at which the L7 requests were processed by this real server. Requests/Sec
Succ_L7_requests Indicates the number of L7 requests that were processed successfully by this real server. Number Ideally, value of this measure should be high.