Collecting and reviewing Cloud Analytics data
Once you have a good sample of data, you will start to see emerging patterns, leading to better understanding of component relationships and interactions that cause performance bottlenecks.
The Cloud Analytics tool is located at the bottom of the Machine details page of your SmartMachine. Once you have a SmartMachine in production, you can begin building instrumentations that allow you to monitor specific aspects of your system infrastructure.
- Click the Metric drop-down menu and select a metric to use for monitoring.
- Click the Decomposition drop-down menu and select how you want to breakdown performance data.
In some cases, you can include a second decomposition
- Click the Create button.
If your system is under load, you will start to see data scroll through the Instrumentation window.
In addition to monitoring a single machine, you can also build instrumentations that monitor the entire cloud.
- On the my.joyent.com portal, click the Analytics tab at the top of the page. This opens the Analytics page.
The instrumentation controls on this page are identical to instrumentation controls available through the Machine Details page. The only difference is you need to specify which cloud you want to monitor.
- Select a cloud you would like to monitor from the Create an Instrumentation on drop-down menu.
- Select a metric from the Metric drop-down menu.
- Select a decomposition from the Decomposition drop-down menu.
Create a predicate to filter the performance data. Selecting a decomposition and building a predicate are optional
- Click the Create button.
If the cloud is under load, you should start to see performance data scroll through the instrumentation window.
Cloud Analytics instrumentations provide various controls for understanding the data collected from your system infrastructure. Both line graphs and heat maps support a basic set of instrumentation controls.
Heat maps provide additional controls for digging down into performance data. Like the above image, the below heat map demonstrates Cloud Analytics capturing data on how CPU threads are distributed. However, this heat maps uses two decompositions. The data collection context is running applications and the sub-second offset. The subsecond offset represents activity that occurs between a measured period of time. So, if the x-axis represents one second intervals, the y-axis represents what occurs within that one second of time.
In the above image, the majority of thread distribution occurs with tar. In the below image, tar has been isolated from the heat map.
|These buttons control your view of data. Click the Play button to pause or resume the scrolling of data. Use the directional buttons to browse forward or backward through data.|
|These buttons control the size of the data sample in view. For example, decreasing the view will display trends that occur over a longer period of time.|
|Element list||This pane displays the items that you are measuring. For example, if you're measuring HTTP server operations decomposed by URL, this pane contains a list of URLs. In a heat map, selecting an item from this list highlights all the buckets that contain measurements for that element. You can use the filters at the bottom of the list to isolate or exclude the selected items.|
|metric||This pane captures the bulk of Cloud Analytics data. You can think of the data captured here as a collection of data "buckets." Each bucket represents an individual datapoint.|
|Bucket list||This pane captures detail about individual datapoints in the data you see. In the above line graph, this pane points out that an application (tar) is processing 99 threads during the one second of time the data was collected.|
The below instrumentation features are only supported by heat maps.
|Granularity||This slider controls the range of values for each bucket.|
|Scale||This slider controls the range of values for the y-axis. The range is given in the text below the control.|
|When the sliders are all the way to the top and the bottom, the scale of the y-axis changes depending on the range of values displayed. If you use the sliders to set an upper or lower limit, the scale doesn't change.|
|Rank and linear|
|This control lets you choose how the density of the buckets in a heat map is rendered.|
Linear coloring means that each bucket is colored according to its value. A bucket whose value is 100 will be 100 times darker than a bucket whose value is 1. Rank-based coloring sorts all the values and distributes the color density among them. Rank-based coloring is particularly useful for discovering outliers. |