Isolating local performance variation in website monitoring
View Patent ↗Systems and methods are provided for isolating performance variation in website monitoring. A set of geographically diverse monitoring agents are identified and a subset of baseline agents is selected from these monitoring agents. The baseline agents are those agents that are geographically close to the target website or otherwise enjoy reliable communications with the target website and are therefore less affected by network latencies. The frequency at which a target website is monitored is weighted in favor of the baseline agents in order to get an accurate baseline metric for the target website while also obtaining information regarding global accessibility of the website from the geographically disperse monitoring agents. The variations in sample frequency can be optimized for each particular set of baseline agents. The set of baseline agents can be pre-assigned or dynamically identified based on an analysis of the performance variation from each of the monitoring agents.
1. A computer implemented method for monitoring a target website from a monitoring server that is communicatively coupled with the target website via a data communication network, the method comprising:
dividing a plurality of agents into baseline agents and monitor agents based on a proximity to the target website, wherein each monitor agent is less proximate to the target website than any baseline agent; and
monitoring the target website in a plurality of monitoring rounds that are weighted in favor of the baseline agents by, in each monitoring round, monitoring the target website from each baseline agent and from at least one of the monitor agents.
2. The method of claim 1 , wherein the proximity is a geographical proximity.
3. The method of claim 1 , wherein the proximity is a network latency.
4. The method of claim 1 , wherein the proximity is a packet round trip time.
5. The method of claim 1 , wherein dividing the plurality of agents comprises identifying those agents having the most reliable responses to a series of network communication tests with the target website to determine the baseline agents.
6. The method of claim 5 , wherein the most reliable responses are those responses having the fastest network round trip time.
7. The method of claim 5 , wherein the most reliable responses are those responses having the least performance variation in website load time.
8. The method of claim 1 , further comprising generating a metric for each of the plurality of agents, wherein the metric comprises a comparison of a measured performance of the agent to the historical average performance of the agent, and wherein dividing the plurality of agents into baseline agents and monitor agents based on a proximity to the target website comprises selecting the baseline agents based on the metric.
9. The method of claim 1 , wherein each monitoring round comprises:
monitoring the target website from each baseline agent;
selecting at least one, but not all, of the monitor agents based on a round-robin technique; and
monitoring the target website from the selected monitor agents.
10. A system for monitoring a target website via a data communication network, the system comprising:
at least one hardware processor; and
a controller communicatively coupled with a plurality of agents, wherein the controller, when executed by the at least one hardware processor,
divides the plurality of agents into baseline agents and monitor agents based on a proximity to the target website, wherein each monitor agent is less proximate to the target website than any baseline agent, and
monitors the target website in a plurality of monitoring rounds that are weighted in favor of the baseline agents by, in each monitoring round, monitoring the target website from each baseline agent and from at least one of the monitor agents.
11. The system of claim 10 , wherein the proximity is a geographical proximity.
12. The system of claim 10 , wherein the proximity is a network latency.
13. The system of claim 10 , wherein the proximity is a packet round trip time.
14. The system of claim 10 , wherein the controller divides the plurality of agents by identifying those agents having the most reliable responses to a series of network communication tests with the target website to determine the baseline agents.
15. The system of claim 14 , wherein the most reliable responses are those responses having the fastest network round trip time.
16. The system of claim 14 , wherein the most reliable responses are those responses having the least performance variation in website load time.
17. A system for monitoring a target website from a monitoring server that is communicatively coupled with the target website via a data communication network, the system comprising:
means for dividing a plurality of agents into baseline agents and monitor agents based on a proximity to the target website, wherein each monitor agent is less proximate to the target website than any baseline agent; and
means for monitoring the target website in a plurality of monitoring rounds that are weighted in favor of the baseline agents by, in each monitoring round, monitoring the target website from each baseline agent and from at least one of the monitor agents.
18. The system of claim 17 , wherein the proximity is a geographical proximity.
19. The system of claim 17 , wherein the proximity is a network latency.
20. The system of claim 17 , wherein the proximity is a packet round trip time.
21. The system of claim 17 , wherein the means for dividing a plurality of agents comprises means for performing a series of network communication tests with the target website and means for identifying agents having the most reliable responses to said series of network communication tests to determine the baseline agents.
22. The system of claim 21 , wherein the most reliable responses are those responses having the fastest network round trip time.
23. The system of claim 21 , wherein the most reliable responses are those responses having the least performance variation in website load time.