Detecting uneven load balancing through multi-level outlier detection
The present disclosure is directed towards systems and methods of detecting a cause of anomalous load balancing among a plurality of servers. A device intermediary to a plurality of clients and a plurality of servers collects values of a plurality of counters. The device identifies a server of the plurality of servers that is an outlier. The device can identify a counter of the plurality of counters that is an outlier based on at least a comparison of values of each of the plurality of counters for each of the plurality of servers. The device can provide, responsive to the determination, an indication that a value of the counter is a factor causing the server to have uneven load balancing during the time interval.
1. A method comprising:
identifying, by a device intermediary between a plurality of servers and a plurality of clients, a plurality of counters for each of the plurality of servers having values identifying statistics of the plurality of servers, the device forwarding network traffic from the plurality of clients to at least one server selected from the plurality of servers in accordance with a load balancing scheme;
identifying, by the device based on a comparison of a number of server hits across the plurality of servers using a statistical technique, a server of the plurality of servers that is an outlier server with respect to an amount of network traffic being forwarded to the plurality of servers;
responsive to identifying the server of the plurality of servers, determining, by the device based at least on a comparison of values of the plurality of counters using a statistical outlier algorithm, that a value of a counter of the plurality of counters for the identified server is an outlier counter with respect to other counters of the plurality of counters for the identified server;
outputting, by the device responsive to the determination, an indication that the value of the counter is a contributing factor of the identified server being the outlier counter.
2. The method of claim 1 , wherein the plurality of counters have values identifying statistics of a plurality of virtual servers on the device used to forward network traffic from the plurality of clients to the plurality of servers.
3. The method of claim 1 , wherein the number of server hits across the plurality of servers indicated uneven forwarding of network traffic to the plurality of servers.
4. The method of claim 1 , wherein the indication that the value of the counter corresponds to the contributing factor causing uneven forwarding of network traffic to the identified server.
5. The method of claim 1 , further comprising comparing, by the device, the number of server hits across the plurality of servers during one or more time intervals.
6. The method of claim 1 , further comprising comparing, by the device, values of the plurality of counters during one or more time intervals.
7. The method of claim 1 , further comprising causing, by the device,
display of the indication that the value of the counter corresponds to the contributing factor of the identified server being the outlier server during one or more time intervals.
8. The method of claim 1 , further comprising adjusting, by the device, a configuration of the device to forward network traffic to at least the identified server more evenly.
9. A device comprising:
one or more processors, coupled to memory, and configured to:
identify a plurality of counters for each of a plurality of servers having values identifying statistics of the plurality of servers, the device intermediary between the plurality of servers and a plurality of clients, the device forwarding network traffic from the plurality of clients to at least one server selected from the plurality of servers in accordance with a load balancing scheme;
identify, based on a comparison of a number of server hits across the plurality of servers using a statistical technique, a server of the plurality of servers that is an outlier server with respect to an amount of network traffic being forwarded to each other server of the plurality of servers;
determine, responsive to identifying the server of the plurality of servers, based at least on a comparison of values of the plurality of counters using a statistical outlier algorithm, that a value of a counter of the plurality of counters for the identified server is an outlier counter with respect to other counters of the plurality of counters for the identified server;
output, responsive to the determination, an indication that the value of the counter is a contributing factor of the identified server being the outlier server.
10. The device of claim 9 , wherein the plurality of counters have values identifying statistics of a plurality of virtual servers on the device used to forward network traffic from the plurality of clients to the plurality of servers.
11. The device of claim 9 , wherein the number of server hits across the plurality of servers indicates uneven forwarding of network traffic to the plurality of servers.
12. The device of claim 9 , wherein the indication that the value of the counter corresponds to the contributing factor causing uneven forwarding of network traffic to the identified server.
13. The device of claim 9 , wherein the one or more processors are further configured to compare the number of server hits across the plurality of servers during one or more time intervals.
14. The device of claim 9 , wherein the one or more processors are further configured to compare values of the plurality of counters during one or more time intervals.
15. The device of claim 9 , wherein the one or more processors are further configured to cause display of the indication that the value of the counter corresponds to the contributing factor of the identified server being the outlier server during one or more time intervals.
16. The device of claim 9 , wherein the one or more processors are further configured to adjust a configuration of the device to forward network traffic to at least the identified server more evenly.
17. A non-transitory computer readable medium storing program instructions for causing one or more processors to:
identify a plurality of counters for each of a plurality of servers having values identifying statistics of the plurality of servers, the one or more processors of a device intermediary between the plurality of servers and a plurality of clients, the device forwarding network traffic from the plurality of clients to at least one selected from the plurality of servers in accordance with a load balancing scheme;
identify, based on a comparison of a number of server hits across the plurality of servers using a statistical technique, a server of the plurality of servers that is an outlier server with respect to an amount of network traffic being forwarded to each other server of the plurality of servers;
determine, responsive to identifying the server of the plurality of servers, based at least on a comparison of values of the plurality of counters using a statistical outlier algorithm, that a value of a counter of the plurality of counters for the identified server is an outlier counter with respect to other counters of the plurality of counters for the identified server;
output, responsive to the determination, an indication that the value of the counter is a contributing factor of the identified server being the outlier.
18. The non-transitory computer readable medium of claim 17 , wherein the plurality of counters have values identifying statistics of a plurality of virtual servers on the device used to forward network traffic from the plurality of clients to the plurality of servers.
19. The non-transitory computer readable medium of claim 17 , wherein the number of server hits across the plurality of servers indicates uneven forwarding of network traffic to the plurality of servers.
20. The non-transitory computer readable medium of claim 17 , wherein the program instructions further cause the one or more processors to adjust a configuration of the device to forward network traffic to at least the identified server more evenly.