Overhead management for event tracing

• US 9,137,136 B2
• Filed: 02/03/2012
• Issued: 09/15/2015
• Est. Priority Date: 02/04/2011
• Status: Active Grant

• Current Assignee: Riverbed Technology, LLC (Riverbed Technology Incorporated)
• Original Assignee: Riverbed Technology Incorporated
• Inventors: Malloy, Patrick, Crosby, Peter Anthony, Meagher, Robert
• Primary Examiner(s): Dalencourt, Yves
• Application Number:
  US13/365,496
• Publication Number:
  US 20130145015A1
• Time in Patent Office: 1,320 Days
• Field of Search: 709/224
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 11/3495 for systems
• CPC Class Codes: G06F 2201/86 Event-based monitoring
• CPC Class Codes: G06F 2201/87 Monitoring of transactions
• CPC Class Codes: G06F 2209/508 Monitor
• CPC Class Codes: H04L 43/14 using software, i.e. software packages network security related monitoring H04L63/1408
• CPC Class Codes: H04L 67/025 for remote control or remote monitoring of applications

Abstract

Embodiments provide methods for managing data generated by software transactions, such as event tracing software. In one embodiment, data generated by event tracing software is monitored. The throughput of the data generated may then be modulated based on various criteria, such as a target data rate. The throughput target may be specified on a per-system basis or individual basis. Based on the throughput, the level of detail recorded is modulated. Individual processes may determine a limit or quota depending on their contribution to the throughput. In one embodiment, the method calls for a trace are modified with different property specifications to meet a desired throughput of event tracing data.

Claims

1. A method of modulating an amount of a resource used by a trace process running on an application server to record transaction data, the method comprising:

receiving a target data rate, wherein the target data rate is a rate at which the transaction data is desired to be recorded by the trace process, and wherein the target data rate is specified as an amount of data per unit time;

determining, by a processor, a quota for each application process in a plurality of application processes based on the target data rate, wherein determining the quota comprises individually setting different quota rates for different application processes based on at least a type and duration of each application process; and

for each application process in the plurality of application processes,

determining, by the processor, the amount of the resource used by the trace process to record transaction data associated with the application process; and

modulating, by the processor, a level of detail written by the trace process based on the amount of the resource used by the trace process in relation to the quota associated with the application process.

2. The method of claim 1, wherein receiving the target data rate comprises receiving a requested data rate from a user.

3. The method of claim 1, wherein determining the quota for the application process in the plurality of application processes comprises determining the quota for a number of bytes written to a storage.

4. The method of claim 1, wherein determining the quota for the application process in the plurality of application processes comprises determining the quota for a number of processor cycles.

5. The method of claim 1, wherein determining the quota for the application process in the plurality of application processes comprises determining a maximum amount of storage space that may be used by the trace process.

6. The method of claim 1, wherein modulating the trace process comprises modulating method calls traced by the trace process.

7. The method of claim 1, wherein modulating the trace process comprises tracing method calls traced by the trace process having a minimum call duration.

8. The method of claim 1, wherein modulating the trace process comprises tracing method calls traced by the trace process having a set of selected properties.

9. The method of claim 1, further comprising:

running a plurality of application processes comprising the trace process; and

determining a communal data rate for the plurality of application processes based on the received target data rate,

wherein determining the quota further comprises determining a corresponding quota for each application process in the plurality of application processes based on the determined communal data rate.

10. The method of claim 9, wherein determining the quota further comprises evenly dividing the communal data rate among the plurality of application processes.

11. An application server configured to trace a transaction serviced by the application server and modulate a resource consumed by tracing the transaction, the application server comprising:

a processor;

an application process, executed by the processor, configured to perform the transaction;

a trace process, executed by the processor, configured to record trace information related to the transaction;

a storage configured to store the trace information recorded by the trace process; and

a throughput manager, executed by the processor, configured to;

receive a requested limit for the trace process, the requested limit being a rate at which the trace information is desired to be recorded to the storage, and wherein the requested limit is specified as an amount of data per unit time,

determine a quota for the application process based on the requested limit, wherein determining the quota comprises individually setting different quota rates for different application processes based on at least a type and duration of each application process, and

modulate a level of detail written by the trace process based on the resource of the application server consumed by the trace process in relation to the quota associated with the application process.

12. The application server of claim 11, wherein the throughput manager is configured to determine the quota for a number of bytes written by the trace process to the storage.

13. The application server of claim 11, wherein the throughput manager is configured to determine the quota for a number processor cycles used to service the trace process.

14. The application server of claim 11, wherein the throughput manager is configured to determine the quota for a maximum amount of storage space that may be used by the trace process.

15. The application server of claim 11, wherein the throughput manager is configured to modulate method calls traced by the trace process.

16. The application server of claim 11, wherein the throughput manager is configured to trace method calls having a minimum call duration.

17. The application server of claim 11, wherein the throughput manager is configured to trace method calls having a set of selected properties.

18. A method for modulating an amount of data recorded for a transaction on an application server, the method comprising:

running a plurality of application processes, one of the plurality of application processes executing the transaction;

monitoring a data activity rate, wherein the data activity rate is a rate at which the data for the transaction is recorded, wherein the data activity rate is specified as an amount of data per unit time, and wherein said monitoring comprises;

removing a token from a token bucket each time a resource is used during a fixed period, and

adding a token to the token bucket each time a resource is not used during a fixed period;

determining, by a processor of the application server, a throughput level based on the monitored data activity rate, wherein the throughput level is associated with a minimum call duration and a set of property types, and wherein the throughput level is computed based on a number of tokens currently in the token bucket and an initial number of tokens that were present in the token bucket;

determining a quota for each application process in the plurality of application processes based on the determined throughput level, wherein determining the quota comprises individually setting different quota rates for different application processes based on at least a type and duration of each application process;

selecting, by the processor, a method call and a property based on the determined throughput level; and

recording, by the processor, the selected method call and the selected property based on the throughput level.

19. The method of claim 18, further comprising:

increasing a level of detail of the recording based on the monitored data activity rate being less than the determined throughput level; and

decreasing the level of detail of the recording based on the monitored data activity rate being greater than the determined throughput level.