|
The Evident ClearStone Live (ECSL) ChronoGraph feature is an optional and separately licensed feature of ECSL. This documentation only applies to ECSL Server deployment for which this feature has been properly licensed and deployed.
The Evident ClearStone ChronoGraph feature is an extension of the Evident ClearStone Live (ECSL) product that offers historical analytics. With the ChronoGraph feature, users can view and analyze historical performance data produced by ECSL systems. The historical data can be trended from hours to months with a granularity of 5-minute to daily intervals. Users can choose from dozens of metrics collected by ECSL. Users can compare current performance metrics against historical averages.
The ChronoGraph feature set includes four main components:
- The ChronoGraph Builder is a designer tool that enables users to design new historical visualizations known as ChronoGraphs.
- The ChronoGraph Viewer is a read-only view tool that enables users to access and view multiple ChronoGraphs.
- The Oracle database used for storing the historical data is known as ClearBase.
- The Database Publisher is a server side component that is responsible for storing the real-time data from the Evident ClearStone OpCache into ClearBase.
Note: To access this feature, users must have a valid ECSL license with the ChronoGraph option enabled. This feature set must be chosen at install time of ECSL. Once installed, the administrator must properly configure the ECSL with the necessary database information in order to publish data into the database for the ChronoGraphs. Please consult the ECSL Installation Guide for further details.
While reading this documentation, there are two user roles to be aware of:
The ChronoGraph Builder is the designer tool used to create new historical visualizations known as ChronoGraphs. Using the Builder, the designer can create templates that display trend charts for a specific performance metric that is stored in ClearBase. Each ChronoGraph is highly customizable. A designer can change the chart type and formatting of the chart, such as axis, titles, legend, etc. Furthermore, the designer can choose from many performance metrics with different levels of granularity. Common aggregation methods are available to consolidate/summarize data across multiple data series or time series. Designers can also specify runtime and default parameters for each ChronoGraph. The parameters allow end users to report on different series of data based on different parametric values.
The Builder user interface consists of the following areas:
- Template Catalog
- Navigation Panel
- Builder Panel
- Preview Panel
- Source Panel
Template Catalog
Each ChronoGraph is an XML template. All the templates are organized in a template catalog where end users can access the published templates. The templates and catalog are stored locally on the ECSL server, where the Builder is installed and accessed. Templates can be organized in a hierarchical grouping. Designers can define one or more groups and levels of groups to store the templates.
The Builder enables designers to create, modify, copy, save a draft, publish, rename, and remove a template. As a template is developed, the designers can preview the results before saving it. A template can be saved in two ways: Publish or Save Draft.Publishing will save the template and make it available for the end users to view. Templates can be saved as drafts preventing end users from accessing a template that is undergoing construction.
Designers can provide meaningful details about the ChronoGraph in the description field. This usually helps end users to understand the information displayed for the ChronoGraph.
Navigation Panel
As templates are created, the navigation panel guides a designer through specific areas of the ChronoGraph to configure. There are six major facets to a ChronoGraph. These are known as “hot spots.”
- Visualization
- Primary Y-Axis
- Secondary Y-Axis (optional)
- X-Axis
- Parameters
- Title & Legend
The input for each hot spot will be displayed in the Builder Panel. Each hot spot will have different properties that can be configured. Some options in the Visualization hot spot may enable/disable other hot spots for the ChronoGraph. For example, the Secondary Y-Axis will only appear if that Overlay Options is enabled in the Visualization hot spot.
As each hot spot is configured for a ChronoGraph, the navigation panel will mark it as complete. Designers can use the navigation panel to go back to update previously configured hot spots. Any hot spot that is disabled will not require any input. A template is only completed when all the required hot spots are completed. At any time, a designer can perform a save draft to store the template.
Builder Panel
The Builder panel will be the primary data input area for the Builder. The contents of the panel will depend on the current active hot spot. Before applying or proceeding to the next hot spot, the Builder will validate that all the required input has been specified for the template. Let’s review the properties of each hot spot.
The navigation buttons at the bottom if the builder panel and their function are:
- Previous – Displays a previous hot spot
- Apply - Updates the current hot spot with the currently configured values (this is not the same as saving)
- Cancel – Discards any changes made to the current hot spot
- Next – Advances to the next hot spot
Visualization
This is the first hot spot for all the ChronoGraphs. Designers decide how the main visualization should appear. The available types of visualizations are mainly charts (i.e. bar, line, area). There may be additional subtypes for each chart type.
For the Cartesian charts like bar, line, and area, these require configuring the X-axis and at least one Y-axis. These charts support time-series based data for a single metric.
The Overlay Options in the Visualization Hot Spot allow designers to apply one of the two following options:
- Include a secondary Y-axis to chart a different, additional metric. This is useful for comparing performance trends of two different metrics. If this overlay option is selected, the designer will need to configure the Secondary Y-axis hot spot.
- Include a different time series for the same metric. This is useful for comparing performance trends of the same metric between two periods of time. The two different time series must share the same starting point and have the same duration. If this overlay option is selected, the designer will need to configure the Time Series #2 area of the X-axis hot spot.
To advance to the next hot spot, click on the Next Button. To only update the current hot spot, click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the Builder would load the Primary Y-Axis hot spot.
Primary Y-Axis
The next hot spot is the Primary Y-axis hot spot. In this hot spot designers would:
- Choose a specific KPI from the available datasets in the data warehouse
- Configure aggregations and grouping
- Format the Y-axis
The KPIs are grouped by family and then by record type. For an Oracle Coherence environment, there would a single family called “Coherence”. The record types are Cluster Statistics, Cache Statistics, Node Statistics, Proxy Node Statistics, and Extend-TCP Session Statistics. Each record type includes a set of KPIs to choose from. These KPIs are already aggregated in the data warehouse for 5 minute, hourly and daily time periods. A single KPI may have different derivations of a calculated value (i.e. average, min, max, sum). Some KPIs have designated units of measurement that can be used for labeling and scaling the chart data.
For a selected KPI, it’s possible the query results produced from the data warehouse may yield multiple unique series of data for the same period of time. By default none of the series will be aggregated. Each series will be treated as distinct rows of data. To aggregate the data across multiple series, designers can define the how data should be grouped and aggregated. The “Group By” fields are based on the non-numeric identifiers of the selected record type. For example, if the record type is “Cache Statistics”, designers can group by Cluster Name, Cache Type, Service Type, Cache Name and Tier. Some record types may require a finer grouping; therefore a second level of grouping can be included. Once the records are grouped, the designer must select a method of aggregation for the metric. This can be:
- Average – compute the average value across all the series based on the specified grouping for each time interval
- Total – sum up values across all the series based on the specified grouping for each time interval
- Min – choose the smallest value across all the series based on the specified grouping for each time interval
- Max – choose the largest value across all the series based on the specified grouping for each time interval
As an example, let’s assume that Cluster_1 has two caches, CacheA and Cache B. Cluster_2 also has two caches, CacheA and CacheB. Let’s also assume that grouping was configured with one identifier, that is, Cache Name and the aggregation was sum. If the resulting dataset included data for all caches on both clusters, than the metrics for CacheA on Cluster_1 will be added to the metrics for CacheA on Cluster_2 since the names of those caches on those two clusters are exactly the same. CacheB metrics would also be summed in this way.
But if the grouping was configured with two identifiers, that is, Cluster Name and Cache Name, than the metrics for CacheA on Cluster_1 would not be added to the metrics for CacheA on Cluster_2 since the cluster identifier for those two data series are different and therefore cannot be grouped. The same goes for CacheB.
In the Y-axis formatting section, the title is the same as the name of the KPI by default. However, the designer can choose to specify a user-defined title for the Y-axis. The scaling and precision of the Y-axis can also be controlled. Lastly, the chart can be rendered with colored gridlines and tick marks on the axis. The designer can choose from a wide range of colors from a palette.
To advance to the next hot spot, click on the Next Button. To only update the current hot spot, click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the Builder would load the Secondary Y-Axis or X-axis hot spot.
Secondary Y-Axis
The Secondary Y-Axis hot spot is enabled when the template is configured with an overlay option with secondary Y-axis for additional KPI. This hot spot enables a designer to select a secondary KPI to compare against the primary KPI. The second KPI is limited to a performance metrics that belongs in the same record type as the primary KPI.
The KPI Grouping and Axis Formatting provide the same functionality as the primary Y-axis.
To advance to the next hot spot, click on the Next Button. To only update the current hot spot, click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the Builder would load the X-axis hot spot.
X-Axis
The next mandatory hot spot for the Cartesian charts is the X-axis hot spot. The designer defines the scope and interval of the data for the selected KPI(s). Up to two time-series can defined. The second time series is only enabled if the designer enabled the Dual Time Series overlay option in the Visualization hot spot.
For each time series, the designer can choose from a list of predefined scopes (i.e. Today, Yesterday, This Week, Last Week, Last 7 Days, This Month, Last Month, User Defined). The relative scopes are auto computed upon the loading of the ChronoGraph in the Viewer. For the User Defined scope, the designer can specify a default start and end date for the scope.
The X-axis aggregation allows the ChronoGraph to aggregate data in terms of time. Let’s assume that we have a template that is based on hourly data and the date range of data to be included is Jun 29th and 30th. Assuming the data includes metrics for 3 caches, I will get a total of 48 data points (corresponding to the 24 hours in each of the two days) for each of the three caches.
If I apply an X-axis aggregation of Total to this template, the ChronoGraph will aggregate the data into one day (that is, one set of 24 hours) and apply an aggregation (in this case total) to the metrics to calculate the result. Refer to the chart below:
The result section is what would be charted in the template.
The X-axis can be formatted with the proper date and time labels based on the selected interval. The designer can specify the title for the X-axis as well. The axis can be rendered with colored gridlines and tick marks on the axis. The designer can choose from a wide range of colors from a palette.
To advance to the next hot spot, click on the Next Button. To only update the current hot spot, click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the Builder would load the Parameter’s hot spot.
Parameters
All the ChronoGraphs provide the option of prompting end users for data input (aka Parameters). The input parameters are used to search and filter for results in the ChronoGraph database. The input parameters are limited to the non-numeric identifiers of the selected record type. As a rule of thumb, designers should specify parameters for the attributes that are being used as part of the grouping.
To specify an input parameter:
- Choose the parameter
- Update the parameter prompt. By default it will use “Select <parameter name>”
- If the input parameter is required for the ChronoGraph, check the Required checkbox.
- If the parameter value should not be changed by the end user, then uncheck Modifiable. This prevents the end users from overriding the default parameter value specified in the Default Value field.
- For the prompt type, select “List-Dynamic”. This option will auto-generate a list of available parameter values for the parameter. This list may be extended in future releases.
- Specify a default value for the input parameter. This value will be used as the default input value for viewing the ChronoGraph. This can be empty, but when the ChronoGraph is viewed, the user interface will prompt for the value.
- The Max Values field can limit the # of return values in the dynamic list. In the event an end user enters a wildcard (*) or a value that results in a large list, the Max Values can limit the # of entries to displayed.
- If there is multiple input parameters defined, the order in which those parameters appear in the viewer can be set. The up arrow button will move the parameter before the previous one. The down arrow button will move the parameter after the next one.
- To add input parameters, click on the + button.
- To remove input parameters, click on the - button.
To advance to the next hot spot, click on the Next Button. To only update the current hot spot, click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the Builder would load the Title and Legend hot spot.
Title and Legend
The final hot spot is the Title and Legend. Here, the designer can customize the ChronoGraph’s title. By default the Builder will auto-generate a title for the chart that contains the template name, primary KPI name, time series scope, and parameter values.
The designer can update the text content of the title. The designer can also add, change or remove text from the title. There are some pre-defined variables (marked with { }) that are included in the title by default: time_series_1, time_series_2, param_value_X. These are generated by the ChronoGraph Viewer based on the current input for the ChronoGraph. The time_series_X variables will be replaced by the start and/or end date of each time series. The param_value_X variables will be replaced by the user entered parameter value for the ChronoGraph.
Lastly, the legend can be turned off in the ChronoGraph. Upon viewing, no legend will be displayed.
This is the last hot spot. The designed has to click on the Apply button. To revert to the template’s original configuration, click on the Cancel button. Once the data input has been validated, the template is ready to be saved.
Preview
Once a template is complete, the designer can access the Preview panel to view the current template’s results and formatting. It is useful for accessing the data and reviewing the formatting of the visualization to ensure the results are well presented.
The Preview panel also exposes the input parameter fields required for the ChronoGraph. Designers can specify different parameter values to ensure the proper input parameters are used for the ChronoGraph and the expected results are presented. The input parameter field utilizes auto-complete features that will automatically lookup available matching parameter values in the database for the current parameter. The designer or end user can choose the value(s) they would like to select in for the parameter.
When a template is configured with the 2nd time series overlay option, the Series 2 date prompts will be enabled. This prompt will constrain the end user to select a date range that starts on the same day as the first and has the same duration.
Source View
As stated earlier, the templates are XML documents that will be stored on the ECSL server. The XML source is exposed in the Source View panel. The contents of the Source View cannot be modified. This is only useful for troubleshooting and debugging the templates.
The ChronoGraph Dashboard/Viewer is the end-user user interface that is used to access the published ChronoGraphs. The ChronoGraph Viewer is deployed as a plug-in for the ECSL Real-Time Console. It allows end users to access and view multiple ChronoGraphs in a single view. The Viewer user interface consists of the following areas:
- Template Catalog
- Parameter Panel
- Viewer Panel
Template Catalog
The Viewer’s Template Catalog is only used to access the published templates. It presents the templates in groups that can be expanded. The catalog is used in conjunction with the Viewer Panel to select the ChronoGraphs for loading.
Parameter Panel
As each template is selected for viewing, the user may be required to enter some input parameters values. The parameters displayed in the Parameter Panel are based on the configuration of the selected template. It includes the prompting of the time series date selection and the input parameters. End users are free to change the selection of the parameter values to generate new results.
Viewer Panel
The Viewer Panel is the container where the generated ChronoGraphs are displayed on the screen. The end user can select the grid layout style of the viewer to display one or more ChronoGraphs. To load a template, the user would select a grid within the Viewer Panel and then select a template to load it into the currently selected viewer quadrant. When the Viewer loads the selected ChronoGraph, it will update the Parameter’s Panel with the parameters. It will attempt to use the default values (if any), otherwise it will prompt the end user to specify the input values. Once this information is provided, the Viewer will render the visualization for the ChronoGraph into the selected grid.
|
