Cluster health indicator with dynamic load correlation

• US 20090287720A1
• Filed: 05/15/2008
• Published: 11/19/2009
• Est. Priority Date: 05/15/2008
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Herter, Thomas, Pruscino, Angelo
• Granted Patent: US 8,549,002 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 11/3006 where the computing system is distributed, e.g. networked systems, clusters, multiprocessor systems multiprogramming arrangements G06F9/46; allocation of resources G06F9/50
• CPC Class Codes: G06F 11/3082 the data filtering being achieved by aggregating or compressing the monitored data
• CPC Class Codes: G06F 11/328 Computer systems status display G06F11/327 takes precedence
• CPC Class Codes: G06F 11/3433 for load management allocation of a server based on load conditions G06F9/505; load rebalancing G06F9/5083; redistributing the load in a network by a load balancer H04L67/1029

Abstract

Systems, methods, and other embodiments associated with producing a proximity display of correlated load metrics associated with members of a cluster are described. One example method includes acquiring metrics data (e.g., load data) from nodes in a cluster. The example method may also include determining a cluster element state based on the metrics data and determining relationships between members of the set of related cluster elements. The method may also include identifying element metric representations for cluster elements based on cluster element states and identifying locations on a proximity display at which element metric representations are to be displayed. The locations may depend on relationships between cluster element states. The method may also include displaying element metric representations at the computed locations to produce a proximity display of correlated load metrics.

Claims

1. A computer-readable medium storing computer-executable instructions that when executed by a computer cause the computer to perform a method, the method comprising:

acquiring sets of metrics data from a set of related cluster elements;

determining a cluster element state based on a set of metrics data associated with the cluster element;

determining a relationship of the cluster element state to the cluster elements states of other members of the set of related cluster elements;

identifying a cluster element metric representation based, at least in part, on the cluster element state;

identifying a location on a proximity display of correlated metrics at which the cluster element metric representation is to be displayed, where the location depends, at least in part, on the relationship of the cluster element state to the cluster element states of the other members of the set of related cluster elements; and

displaying the cluster element metric representation at the identified location on the proximity display.

2. The computer-readable medium of claim 1, where a set of metrics data associated with a cluster element identifies a load experienced by the cluster element and where the cluster element is one of, a processor, a memory, a disk, a connection, and a queue.

3. The computer-readable medium of claim 1, where the method includes storing the sets of metrics data in a dependency graph, where a node in the dependency graph stores metrics data from an individual cluster element, and where a branch of nodes in the dependency graph stores metrics data from a related group of cluster elements.

4. The computer-readable medium of claim 3, where membership in the related group of cluster elements is dynamic and depends on a user specified correlation between metrics.

5. The computer-readable medium of claim 1, where determining a cluster element state includes analyzing a set of metrics data associated with the cluster element in light of cluster element metadata associated with the cluster element, where the cluster element metadata includes a set of ranges for members of the set of metrics data, where members of the set of ranges are associated with cluster element states.

6. The computer-readable medium of claim 5, where the set of ranges are dynamic and configurable.

7. The computer-readable medium of claim 1, where the cluster element metric representation is dynamically configurable with respect to one or more of, shape, size, color, color intensity, motion, transition, and transition rate.

8. The computer-readable medium of claim 1, where the method includes re-determining a cluster element state based on metrics data received after a previous cluster element state determination and where the method includes periodically refreshing a cluster element metric representation on the proximity displayed based on the re-determined cluster element state.

9. The computer-readable medium of claim 1, where the method includes producing a set of proximity displays and periodically displaying different members of the set of proximity displays, where members of the set of proximity displays are related to metrics associated with subsets of the metrics data, where members of the proximity displays are associated with one or more of, processor load, disk load, memory load, bandwidth, and queue size.

10. The computer-readable medium of claim 1, where the method includes providing a time series of proximity maps.

11. The computer-readable medium of claim 1, where the method includes displaying a representation of a previous grouping of metric representation elements in conjunction with a current grouping of metric representation elements.

12. The computer-readable medium of claim 1, where the method includes computing a logical center for the proximity display, where the logical center will be the location at which cluster element metric representations associated with a greatest load are to be displayed.

13. The computer-readable medium of claim 1, where computing the logical center of the display depends, at least in part, on an amount of space available to display the proximity display, on a shape of the proximity display, on the number of cluster element metric representations to be displayed, and on the size of the cluster element metric representations to be displayed.

14. The computer-readable medium of claim 1, where the method includes computing two or more logical centers for the proximity display, where a logical center will be the location at which cluster element metric representations with greatest loads associated with two or more users, two or more services, two or more resources, and two or more customers, are to be displayed.

15. The computer-readable medium of claim 1, where the method includes controlling a reader pulse associated with a reader logic that provides members of the sets of metrics data.

16. The computer-readable medium of claim 1, where the method includes controlling a writer pulse associated with a writer logic that provides the proximity display.

17. The computer-readable medium of claim 1, where the method includes receiving a drill through signal from a system on which the proximity display is displayed, where the drill through signal is associated with a cluster element metric representation, and selectively providing a drill through display element to provide data upon which the element metric representation is based.

18. A system, comprising:

a correlation specification logic to receive data concerning a defined correlation between load metrics associated with nodes in a cluster;

a data store to store load data associated with nodes in the cluster, where the load data is acquired in real time from the nodes;

a reader logic to provide the load data;

a display configuration logic to determine node state representations to be displayed on a proximity display, where the node state representations are based, at least in part, on the load data and on the load metrics; and

a display control logic to display a proximity display of correlated load metrics as represented by node state representations, where the location of the representation of node states depends, at least in part, on a defined correlation between load metrics and on a discovered correlation between load metrics.

19. The system of claim 18, where the data store is to organize the load data into a heterogeneous list of node entries, where a node entry includes a heterogeneous list of metrics data.

20. The system of claim 19, where the display configuration logic is to identify an inner region on the proximity display in which a first set of related node state representations are to be displayed.

21. The system of claim 20, where the display configuration logic is to identify an outer region on the proximity display in which a second set of related node state representations are to be displayed, where the first set of related node state representations and the second set of related node state representations are mutually exclusive.

22. The system of claim 21, where the display configuration logic is to identify a field region on the proximity display in which a third set of node state representations are to be displayed, where the first set of related node state representations, the second set of node state representations, and the third set of node state representations are mutually exclusive.

23. The system of claim 18, where the display control logic is to periodically refresh the proximity display based on re-determined node state representation locations that depend, at least in part, on re-determined node state representations.

24. The system of claim 23, where the display control logic is to display a drill-through data upon detecting a user interaction with a portion of the proximity display.

25. A system, comprising:

means for acquiring load data from related nodes in a cluster;

means for computing a set of states for nodes in the cluster and for computing a set of correlations between nodes in the cluster, where a member of the set of correlations depends on one or more of, a defined correlation, and a discovered correlation; and

means for displaying a proximity map of correlated metrics based, at least in part, on the set of correlations.