Systems and methods for gathering deployment state for automated management of software application deployments

• US 7,865,888 B1
• Filed: 12/21/2005
• Issued: 01/04/2011
• Est. Priority Date: 12/21/2004
• Status: Active Grant

• Current Assignee: Citrix Systems, Inc. (Cloud Software Group)
• Original Assignee: Zenprise, Inc. (Cloud Software Group)
• Inventors: Watson, Nathan, Hassan, Tanvir, Roach, Kelly Brian, Qureshi, Waheed
• Primary Examiner(s): Zhen; Wei Y
• Assistant Examiner(s): Brophy; Matthew J
• Application Number:
  US11/316,696
• Time in Patent Office: 1,840 Days
• Field of Search: 717120-135, 717168-178
• US Class Current: 717/168
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06N 5/048 Fuzzy inferencing

Abstract

Systems and methods for obtaining information about a deployment of a software application are disclosed. One system comprises a plurality of deployment servers running components of the software application deployment, a plurality of monitors associated with the deployment servers, and an analysis subsystem that generates requests for state metrics from the deployment, based on rules describing known problems with the software application. The analysis subsystem transmits each request for state metrics to a monitor associated with a deployment server that is associated with the request. The monitors are configured to receive the requests for state metrics from the analysis subsystem and collect the state metrics from one of the components of the software application deployment. The monitors, which can reside on the deployment servers or locally with respect to the analysis subsystem, can transmit the collected state metrics back to the analysis subsystem. An application model storing a representation of objects of the deployment can also be provided, wherein the analysis subsystem queries the application model for deployment information necessary to transmit the requests for state metrics to appropriate ones of the monitors. The monitors can be configured to conduct post-processing of collected state metrics, and also to prevent redundant collection.

Claims

1. A system for obtaining information about a managed deployment of a software application, comprising:

a plurality of deployment platforms, at least some of which are running components of the managed software application deployment;

a meta-application computer system comprising one or more computers, said meta-application system programmed to implement at least;

a discovery component configured to use an initial deployment- independent abstract model of the software application to generate a context- sensitive application model of the managed deployment, by discovering information from the managed deployment and using the discovered information to populate the abstract model;

a plurality of monitors, each monitor associated with a subset of the deployment platforms and configured to gather state metric data from said associated subset; and

an analysis subsystem that generates requests for state metric data for specific state metrics from the managed deployment, based on deployment-independent encoded knowledge or based on deployment-independent encoded remedies associated with the encoded knowledge, the abstract model and the encoded knowledge being generated from analyses of reference materials for the software application being managed and empirical analyses done on a plurality of different deployments of the managed software application, the encoded knowledge describing known problems with the software application, the analysis subsystem configured to transmit each request for state metric data to a monitor associated with one of the deployment platforms that is associated with the request, the analysis subsystem configured to query the application model for deployment information necessary to transmit the requests for state metric data to appropriate ones of the monitors;

wherein each monitor is configured to receive requests for state metric data from the analysis subsystem and collect state metric data responsive to the requests from the platform subset associated with the monitor;

wherein at least some of the monitors are each configured to run a plurality of probes corresponding to multiple state metrics, each probe configured to gather data for a specific one of the state metrics;

wherein the meta-application system is configured to receive updates to the encoded knowledge and/or the abstract model from one or more update servers.

2. The system of claim 1, wherein at least some of the monitors reside on the deployment platforms.

3. The system of claim 1, wherein at least one of the monitors resides locally with respect to the analysis subsystem.

4. The system of claim 1, wherein the monitors transmit at least some of the collected state metric data to the analysis subsystem.

5. The system of claim 1, wherein the analysis subsystem couples at least some of the requests for state metric data with postprocessors, the monitors configured to use the postprocessors to apply filtering, aggregation, and/or transformation to the collected state metrics metric data to generate separate streams of output data transmitted by the monitors.

6. The system of claim 1, wherein at least one of the monitors is responsive to receipt of two or more substantially identical requests for state metric data from two or more different requesting components of the analysis subsystem by collecting the requested state metric data only once and then transmitting the collected state metric data only once to the analysis subsystem, the analysis subsystem configured to transmit the collected state metric data to the two or more different requesting components.

7. The system of claim 1, wherein the encoded knowledge comprises a plurality of rules, one of the rules specifying a deployment condition associated with a specific state metric, the analysis subsystem configured to vary a rate at which the analysis subsystem requests state metric data for the specific state metric, based upon a degree to which the analysis subsystem has determined that other deployment conditions specified by the rule are met.

8. The system of claim 1, wherein, prior to transmitting each state metric data request, the analysis subsystem:

computes costs of each of a set of the state metric data requests and associated state metrics;

computes contribution values of each of the set of state metrics based on the encoded knowledge; and

selects state metrics for which state metric data is to be requested by balancing the contribution values against the costs.

17. The system of claim 1, wherein at least one of the monitors is configured to collect state metric data responsive to a plurality of the requests, and send the collected state metric data metrics to the analysis subsystem together in a single batch.

18. The system of claim 1, wherein at least some of the deployment platforms comprise servers.

21. The system of claim 1, wherein at least some of the state metric data is numerical.

22. The system of claim 21, wherein at least some of the monitors are configured to gather streams of numerical state metric data values for grouped evaluation by the analysis subsystem.

23. The system of claim 1, wherein the encoded knowledge comprises a collection of rules describing potential problems associated with the software application, at least one of the rules comprising a logical combination of conditions, at least one of the conditions being satisfiable by a feature of the managed deployment.

24. The system of claim 23, wherein:

at least some of the monitors reside on the deployment platforms; and

at least one of the monitors residing on one of the deployment platforms is configured to collect state metric data from said deployment platform and analyze the collected data to identify one of the features or problems of one of the rules of the encoded knowledge.

25. The meta-application system of claim 23, wherein at least one of the rules is configured to be used by the analysis subsystem to construct a virtual software logic circuit comprising:

a plurality of atomic gates each configured to be used by the analysis subsystem to test whether one of said features satisfies one of the conditions of the rule, and to produce an output bead if the condition is satisfied; and

a plurality of operator gates each having one or more inputs, each of the inputs configured to receive output beads from one of the atomic gates or from another one of the operator gates, each operator gate configured to be used by the analysis subsystem to test for a condition associated with received beads and to'"'"'produce an output bead if the condition is met;

wherein the rule is configured to be applied by the analysis subsystem to automatically detect one of the problems within the deployment based on output beads generated from use of the operator gates.

26. The system of claim 1, wherein the monitors are configured to execute the remedies upon the deployment platforms, in order to counter problems detected by the analysis subsystem using the encoded knowledge.

27. The system of claim 1, wherein the encoded knowledge has metadata information mapping said state metrics to said monitors.

28. The system of claim 1, wherein at least some of the monitors are configured to load new probes for gathering state metric data for new state metrics at runtime.

29. The system of claim 1, further comprising a telemetry component and a telemetry priority queue (TPQ), wherein the monitors are configured to send the collected state metric data to the telemetry component as a plurality of telemetry data packets representing telemetry-processing tasks, the telemetry component configured to assign a priority value to each data packet based on time sensitivity, the TPQ configured to sequence the telemetry-processing tasks based on the priority values assigned by the telemetry component to the data packets.

30. The system of claim 29, further comprising a scheduled priority queue (SPQ) of scheduled tasks each having a priority value, the priority value of each scheduled task being zero before a time at which the task is scheduled to be executed and increasing over time after the scheduled time of execution, the SPQ configured to sequence the scheduled tasks based on the priority values of the scheduled tasks, wherein the priority values of the scheduled tasks are in the same units as the priority values of the telemetry-processing tasks, the meta-application system configured to perform the telemetry-processing tasks and the scheduled tasks in a combined sequence in which higher priority value tasks are performed before lower priority value tasks.

31. The system of claim 1, wherein at least one of the monitors resides locally with respect to the analysis subsystem.

32. The system of claim 1, wherein the meta-application system is configured to archive the application model.

33. The system of claim 32, wherein the discovery component is configured to store the application model in an SQL database that can be interrogated using SQL queries.

34. The system of claim 1, wherein the meta-application system is configured to choose, after the discovery component generates the application model, (1) an initial set of deployment state metrics to observe, (2) initially acceptable values for deployment parameters including at least one of CPU utilization, volume of data traffic, and disk utilization, and (3) schedules for maintenance tasks to be conducted on the deployment by the meta-application system.

35. The system of claim 1, wherein the application model comprises a graph of objects each representing an element of the managed deployment, the application model classifying relationships between the objects into at least five types, including dependency relationships, delegation relationships, containment relationships, policy propagation relationships, and event relationships.

36. The system of claim 1, wherein the discovery component is configured to select an initial set of monitors for use with the managed deployment.

37. The system of claim 1, wherein the abstract model is provided with an uninstalled version of the meta-application.

38. The system of claim 1, wherein:

at least some of the monitors reside on the deployment platforms;

at least some of the monitors are configured to load new probes for gathering state metric data for new state metrics at runtime;

at least some of the state metric data is numerical;

at least some of the monitors are configured to gather streams of numerical state metric data values for grouped evaluation by the analysis subsystem;

the encoded knowledge comprises a collection of rules describing potential problems associated with the software application, at least one of the rules comprising a logical combination of conditions, at least one of the conditions being satisfiable by a feature of the managed deployment;

the encoded knowledge has metadata information mapping said state metrics to said monitors;

at least one of the monitors residing on one of the deployment platforms is configured to collect state metric data from said deployment platform and analyze the collected data to identify one of the features or problems of one of the rules of the encoded knowledge; and

the meta-application system further comprises a telemetry component, the monitors being configured to send the collected state metric data to the telemetry component as a plurality of data packets, the telemetry component configured to assign a priority value to each data packet, the analysis subsystem configured to analyze the data packets in a sequence that is based on the priority values assigned by the telemetry component to the data packets.

39. The system of claim 1, wherein:

at least some of the monitors reside on the deployment platforms;

at least some of the monitors are configured to load new probes for gathering state metric data for new state metrics at runtime;

at least some of the state metric data is numerical;

at least some of the monitors are configured to gather streams of numerical state metric data values for grouped evaluation by the analysis subsystem;

the encoded knowledge comprises a collection of rules describing problems associated with the software application, at least one of the rules comprising a logical combination of conditions individually satisfied by features of the managed deployment;

the analysis subsystem includes a plurality of feature detectors configured to analyze the state metric data collected by the monitors to generate said features;

the encoded knowledge has metadata information mapping said features to the feature detectors, and/or to new probes for gathering state metric data from which the feature detectors can derive the features; and

the meta-application system further comprises a telemetry component, the monitors being configured to send the collected state metric data to the telemetry component as a plurality of data packets, the telemetry component configured to assign a priority value to each data packet, the analysis subsystem configured to analyze the data packets in a sequence that is based on the priority values assigned by the telemetry component to the data packets.

9. A method of obtaining information about a deployment of a software application, comprising:

generating a context-sensitive application model of the deployment by discovering information from the deployment and using the discovered information to populate a deployment-independent abstract model of the software application;

using the application model to choose and activate a set of monitors for gathering state metric data from the deployment, each monitor associated with one of a plurality of deployment servers, at least some of the servers running, components of the software application, each monitor configured to gather state metric data for deployment state metrics from the server associated with the monitor;

generating requests for state metric data for specific deployment state metrics from the deployment, based on deployment-independent encoded rules describing known problems with the software application or based on deployment-independent encoded remedies associated with the problems; and

accessing the application model to obtain deployment information necessary to transmit the requests for state metrics to appropriate ones of the monitors; and

using the deployment information obtained from the application model to transmit each request for state metric data to an appropriate one of the monitors;

wherein each monitor is configured to receive requests for state metric data from the analysis subsystem and collect state metric data responsive to the requests from the server associated with the monitor;

wherein the method is performed by one or more computing devices.

10. The method of claim 9, wherein at least some of the monitors reside on the deployment servers.

11. The method of claim 9, wherein at least one of the monitors resides locally with respect to a component generating the requests.

12. The method of claim 9, further comprising receiving at least some of the collected state metric data metrics from the monitors.

13. The method of claim 9, further comprising:

coupling at least some of the requests for state metric data with postprocessors, the monitors configured to use the postprocessors to apply filtering, aggregation, and/or transformation to the collected state metric data to generate separate streams of output data; and

receiving the separate streams of output data from the monitors.

14. The method of claim 9, wherein generating the requests comprises generating two or more substantially identical requests for state metric data, at least one of the monitors being responsive to receipt of the two or more substantially identical requests for state metric data metrics by collecting the requested state metric data only once, the method further comprising:

receiving the collected state metric data only once from the monitor; and

transmitting the collected state metric data to different requesting components.

15. The method of claim 9, wherein one of the rules specifies a deployment condition associated with a specific state metric, the method further comprising varying a rate at which state metric data for the specific state metric is requested, based upon a degree to which other deployment conditions specified by the rule are met.

16. The method of claim 9, further comprising, prior to transmitting each state metric data request:

computing costs of each of a set of state metric data requests and associated state metrics;

computing contribution values of each of the set of state metrics based on the rules; and

selecting state metrics for which state metric data is to be requested by balancing the contribution values against the costs.

19. A computer program which embodies the method of claim 9, represented in computer storage.

40. The method of claim 9, further comprising receiving updates to the encoded knowledge and/or the abstract model from one or more update servers.

20. A system for obtaining information about a deployment of a software application, comprising:

a plurality of deployment platforms, at least some of which are running components of the software application deployment;

a meta-application computer system comprising one or more computers, said meta-application system programmed to implement at least;

a discovery component configured to use an initial deployment-independent abstract model of the software application to generate a context-sensitive application model of the deployment, by discovering information from the deployment and using the discovered information to populate the abstract model, the application model comprising an object graph representation of the deployment, at least some of the objects corresponding to hardware or software elements of the deployment;

a plurality of monitors, each monitor associated with a subset of the more deployment platforms and configured to gather state metric data for one or more deployment state metrics from said associated subset;

an analysis subsystem that generates requests for state metric data for specific state metrics from the deployment, based on root cause analysis algorithms associated with the software application, the analysis subsystem configured to transmit each request for state metric data to a monitor associated with one of the deployment platforms that is associated with the request, the analysis subsystem configured to query the application model for deployment information necessary to transmit the requests for state metric data to appropriate ones of the monitors;

wherein each monitor is configured to receive requests for state metric data from the analysis subsystem and collect state metric data responsive to the requests from the platform subset associated with the monitor;

wherein at least some of the monitors are configured to run a plurality of probes corresponding to multiple state metrics, each probe configured to gather data for a specific one of the state metrics;

wherein at least one of the root cause analysis algorithms comprises;

identifying a plurality of suspect objects of the object graph; and

using pattern-recognition against portions of the object graph having at least some of the suspect objects, to identify objects that are root cause candidates.