Method for qualitative analysis of system performance correlation factors

• US 9,391,866 B1
• Filed: 11/14/2013
• Issued: 07/12/2016
• Est. Priority Date: 11/14/2013
• Status: Active Grant

• Current Assignee: T-Mobile Innovations LLC (Deutsche Telekom AG)
• Original Assignee: Sprint Communications Company LP (Deutsche Telekom AG)
• Inventors: Martin, Justin A., McMahon, Margaret M., Washburn, Brian J.
• Primary Examiner(s): Wang, Liangche A
• Application Number:
  US14/080,245
• Time in Patent Office: 971 Days
• Field of Search: 709/223, 709/224, 709/225, 709/226
• US Class Current: 1/1
• CPC Class Codes: H04L 41/0681 Configuration of triggering conditions
• CPC Class Codes: H04L 43/04 Processing captured monitoring data, e.g. for logfile generation
• CPC Class Codes: H04L 43/0817 by checking functioning
• CPC Class Codes: H04L 43/16 Threshold monitoring

Abstract

A method of maintaining server performance comprises receiving, at an analysis server, server performance metrics from at least one monitored server, wherein server performance metrics comprise information about processor usage, memory usage, and network activity and comparing the received metrics with a baseline of values that was established for the server performance metrics based on a variance analysis of historical values. The method further comprises determining that the value of at least one server performance metric of the monitored server is consistently above an upper limit of a range of acceptable deviation from the baseline, analyzing historical occurrences of similar departures of server performance metrics from the range of acceptable deviation, and generating an alert indicating that a risk pattern has been identified and suggesting shutting down a locked process on the monitored server in order to return the server performance metrics to values within the range of acceptable deviation.

Claims

1. A method of maintaining server performance in a network, comprising:

receiving, at an analysis server of the network, server performance metrics from at least one monitored server, wherein server performance metrics comprise information about processor usage, memory usage, and network activity;

comparing, by the analysis server, the received server performance metrics with a baseline of values that was established for the server performance metrics based on a variance analysis of historical values;

determining, by the analysis server, that the value of at least one server performance metric of the monitored server is consistently above an upper limit of a range of acceptable deviation from the baseline over a defined time, wherein the value of the at least one server performance metric determined to be consistently above the upper limit is a departure from the range of acceptable deviation;

analyzing, by the analysis server, historical occurrences of similar departures of server performance metrics from the range of acceptable deviation over the defined time, wherein the analysis identifies historical occurrences being within a predefined range of similarity to the departure by the at least one monitored server over the defined time;

generating, by the analysis server for a work station, an alert indicating that a risk pattern of reduced server performance has been identified based on the analysis and suggesting shutting down a locked process on the monitored server; and

shutting down, via the work station, the locked process on the monitored server in response to the alert, wherein shutting down the locked process returns the server performance metrics to values within the range of acceptable deviation.

2. The method of claim 1, further comprising:

generating, by the analysis server, a trace of the values of the server performance metrics compared to the baseline and the range of acceptable deviation from the baseline.

3. The method of claim 1, wherein the similar departures that are analyzed are at least 80 percent similar to the current departure from the range of acceptable values.

4. The method of claim 1, wherein the variance analysis is a variance analysis of historical values for each hour of each day of the week, of the one or more metrics of the monitored server during periods of acceptable server performance.

5. The method of claim 4, wherein the server performance metric values for each hour of a day are calculated into the baseline at the end of the day.

6. The method of claim 1, wherein the suggestion to shutting down the locked process is based on a result of the analysis of historical occurrences that indicates that, historically, a locked process has been a reason for the server performance metrics to be consistently above the range of acceptable deviation.

7. The method of claim 1, wherein the range of acceptable deviation is bounded by a upper limit and a lower limit of server performance values that a server reaches without triggering an alert.

8. The method of claim 1, wherein the analyzing, by the analysis server, comprises generating a trace of server performance metrics and determining whether server performance metrics correspond with a pattern of historic performance values within a predefined range of similarity over the defined time.

9. The method of claim 1, further comprising adjusting, by the analysis server, the range of acceptable deviation from the baseline based on at least one of:

analysis of loads on monitored servers, a change in a server performance metric corresponding to a beginning of production workload processing, a decrease in amplitude of the server performance metric while maintaining a magnitude over the defined time, or a server performance metric with an amplitude and magnitude without the range of acceptable deviation but shifted on a time basis.

10. A method of maintaining server performance in a network, comprising:

retrieving, by an analysis server of the network, server performance metrics information of at least one monitored server from a data store, wherein server performance metrics information comprises information about processor usage, memory usage, and network activity;

comparing, by the analysis server, the retrieved server performance metrics information of the at least one monitored server with a baseline of acceptable values that was established for the server performance metrics based on a variance analysis of historical values;

determining, by the analysis server, that the value of at least one server performance metric of the at least one monitored server is steadily below the lower limit of a range of acceptable deviation from the baseline over a defined time and indicates a reduction in processing workload compared to historical processing workloads, wherein the value of the at least one server performance metric determined to be steadily below the lower limit is a departure from the range of acceptable deviation;

analyzing, by the analysis server, historical departures of the server performance metrics from the range of acceptable deviation that are within a predefined range of similarity to the departure over the defined time, wherein the analysis identifies a risk pattern of reduced server performance;

generating, by the analysis server, an alert that is configured to indicate that a risk pattern of reduced server performance has been identified based on the analysis; and

in response to the alert, performing maintenance on a load balancer of the network.

11. The method of claim 10, wherein server performance metrics information contained in the data store is stored in the data store by the monitored server.

12. The method of claim 10, wherein performing maintenance on the load balancer of the network returns the server performance metrics to values within the range of acceptable deviation, and wherein performing maintenance on the load balancer is based on a result of the analysis of historical occurrences that indicates that, historically, a load balancer failure has been a reason for similar departures from the range of acceptable deviation.

13. The method of claim 10, wherein the alert is assigned an urgency value based on the analysis of historical departures.

14. The method of claim 13, wherein the alert is ranked in a queue according to the urgency value assigned.

15. The method of claim 10, wherein the range of acceptable deviation is bounded by an upper limit and a lower limit of server performance values that a server reaches without triggering an alert.

16. The method of claim 10, further comprising adjusting, by the analysis server, the range of acceptable deviation from the baseline based on at least one of:

analysis of loads on monitored servers, a change in a server performance metric corresponding to a beginning of production workload processing, a decrease in amplitude of the server performance metric while maintaining a magnitude over the defined time, or a server performance metric with an amplitude and magnitude without the range of acceptable deviation but shifted on a time basis.

17. The method of claim 10, wherein the analyzing, by the analysis server, comprises generating a trace of server performance metrics and determining whether server performance metrics correspond with a pattern of historic performance values within a predefined range of similarity over the defined time.