What is Real-Time Statistics?
Real-Time Stats (RTS) allow you to analyze your domain’s incoming query traffic in real-time. Aggregate and/or filter query data by:
- Domain
- Resource Record
- Location
- Time Frame
Here are some of the ways you can use RTS, which we will demonstrate in the following tutorials:
- View query data for the week, month, or year
- View domains with the largest number of queries
- Send query logs to 3rd party storage
- Set alerts when query rates increase above historical thresholds
- Download historical query data
- More coming soon!
Terminology
Before we dive into the tutorials, we’d like to define a few terms so that we have a common language when using RTS.
Time Series: A set of points that count as a function of time. In the initial download in chart view, each time series is represented by a set of queries for a particular location because that data has been rolled up by record name and record type.
Bucket: When dimensions have specific values, they become a bucket (a collection of dimensions). Each bucket becomes a plot on the graph.
Dimensions: Location, record type, and/or record name are the “dimensions” of a bucket.
Roll-Up By: This is a term we will use a lot in Chart View. You can “roll-up” the different dimensions to remove them from the chart by checking the box next to that dimension. For instance, if you roll up all of the dimensions except for “Location”, then the chart will show the number of queries per location as a function of time.
Legend of Symbols
◊ = Root record
* = Wild card record name
# = NX domain record name
∑ = Aggregated value in the table
Chart View
When you first open RTS, the initial load will be in Chart View. The Chart View Dashboard will display the top five-time series: the number of queries your domain has received over a 5 day period.
This initial download will provide you with a quick overview of all of your queries by location. Any abnormal spikes will appear here, letting you know as soon as possible if there are regional issues that need to be resolved.
Each time-series in the chart shows the “time history” of queries received at that particular location. Time series are color-coded by location (PoP) within our network. You can see which color represents which location in the legend below the chart.
Because domains incur so many queries even within just a few minutes, each data point within a time-series represents the number of queries your domain received at that location over a 30-minute interval.
Date Range
The date range control allows you to select how much historical data will load. RTS stores your historical data as far back as 30-days.
Time Series
The Time Series drop-down allows you to select the number of time series to display in the Chart. Time series are sorted by the most number of queries received with the time frame.
Filters
Filters allow you to select and view only data that matches the filter you have selected. For example, if we click on the Location checkbox, and then select San Jose from the drop-down, only queries from San Jose will be displayed. Or if you click on the Record Type checkbox, and then select the Record Type, A, from the drop-down, only queries from San Jose for A records will be displayed.
You can combine multiple filters, but only one type of each filter can be chosen at a time. For example, you can Filter by Location: Ashburn, VA and Record Type: A Records. But you cannot filter by Location: Ashburn, VA, and Location: Seattle, WA.
Roll-up by
The “Roll-Up By” feature allows you to combine time series. For example, if you click the “All” checkbox, then all the Time series will be combined and the Chart will show the total number of queries, as a function of time, that your domain received over your selected time frame. If you’re following along, please unselect the “All” checkbox to show all the Time series.
You can also combine multiple time series together by selecting the dimensions you want to see together. For example: leave the “Record Type” and “Record Name” checkboxes selected.
Now all the record types and record names will be combined, and the remaining Time series will only have location dimensions. Effectively, you are now seeing the queries received at each of the top five DNS Made Easy name server locations. The legend at the bottom of the Chart will show the series only by location.
This is the same view you will remember from the initial download. That’s because this view is quick to load and provides a general overview of your domain’s query traffic.
Suppose that we want to see which records (Record Names) were queried at each location. Simply unselect the “Record Type” checkbox so that the roll-up is only by record type.
Unselecting the “Record Type” checkbox requires Real-time stats to download more detailed data. Depending on the number of records composing your domain, this download can take 10 to 20 seconds or more. Once the data is loaded, the Chart will display the Time series for locations and record names.
We’ve seen how to display the total queries for your domain, the queries at each location, and the queries for record names at each location.
Suppose that we would like to display the queries for your domain’s record names. If we select the “Location” checkbox in the “Roll-Up By” section, what remains in the Chart are the time series for each record name in your domain.
How to Interact with the Chart
Queries
If you mouse-over a data point of a time series, a pop-up window will display the start of the aggregation time interval. The pop-up will also display the number of queries for that time series received in that time interval. Notice that when you mouse over a data point, the mouse cursor changes from a double-sided arrow to a hand.
Highlighting
When you hover over an item in the legend, it will highlight the associated time-series and fade out the rest. This is really helpful if you are trying to troubleshoot specific issues by location or other filters.
Hiding
We’ve just seen that by mousing over the name of a time series in the legend, we can highlight that time series. If we instead click on the name of the time series in the legend, then that time series will be hidden from the Chart. You can hide as many time series as you wish. To display a hidden time series, click it again.
Zooming
For this function, you will need a mouse with a scroll wheel. If you do, then move the mouse cursor onto the Chart. When the mouse cursor changes to a double-sided arrow, you can turn your mouse wheel (our make the scroll gesture on a trackpad or magic mouse), the Chart will zoom around the cursor’s location. By placing the mouse cursor over areas of interest, you can zoom in and out of any time on the Chart.
Tabular View
The “Tabular Data” tab will open with a summary table that displays all of the time series. Each time series will show the total number of queries, as well as the minimum and a maximum number of queries received during any of the aggregation intervals.
NOTE: If you wish to view aggregated data in the Tabular data tab, please first perform your desired roll-ups on the chart tab.
Modify Display
The table will initially show only a 15-time series, sorted by the total number of queries. You can adjust this number by clicking the drop-down list on the top left-hand side of the table. You can choose to show: 15, 50, 250, 1000, and All the time-series in the table.
Paging Through Time Series
On the bottom left-hand side of the table, you should see the index range of the currently displayed time series, and the total number of time series the application has loaded. The control on the bottom right-hand side of the table will help you navigate through the time series one page at a time. If you click on the “Next” button, the table will show the next page of the time series. Clicking on “Previous” will cause the table to show the previous set of time series, if available. Clicking on any of the numbers will cause the table to display that page of time series.
Search
The text box on the top right-hand side of the table, labeled “Search:” provides a convenient method for filtering time series. For example, if you type “AAAA”, then only time series that have “AAAA” text in the values of their dimensions (i.e. location, record type, or record name containing an AAAA) will be displayed.
If you would like to search for more than one word, you can type a space between the search words. Only time series that have all the space-separated words will be displayed. For example, if we type “AAAA ns10″ into the search box, then the table will only display the only 13 time series that contain both “AAAA” and “ns10″.
Filters
The table-level search is very useful for quick searches, but suppose that you want to view the total queries for each location. Recall from our earlier discussion that to view the total queries received at all locations means that the other two dimensions (Record Name and Record Type) have to be rolled-up.
You may have noticed a text box at the bottom of each column. For example, the “Location” column has a text box at the bottom of its column that holds the default text “Search Location”. In our example, we will type the “∑” symbol into the search text boxes at the bottom of the “Record Name” and “Record Type” columns. And now we see each “Location” and the total number of queries at that location.
The “∑” symbol displayed in a table column means that the dimension has been “Rolled-up By”. When all the search terms are removed, you may notice that the first time series in the table has the ∑ symbol in each of the three dimensions: location, record type, and record name. This time series represents the total of all the time series. Recall from our previous discussion that when we roll-up by all the dimensions, we are combining all of the time series. Interestingly, the “Total Queries” for this time series is the total number of queries your domain received during the requested time-window.)
View Details by Time Series
Suppose that we have filtered our time series so that we are displaying the queries for each location.
But now we would like to see the individual data points for the time series representing all the queries received by our Hong Kong name server. Click on the green “+” symbol next to the location name. This will open a chart of that time series, and below it, a table that shows the number of queries received in each of the aggregation intervals.
This sub-table provides the same navigation and paging behavior as the main table.