Charts

Line Chart Resolution

Line chart resolution simply means how many points of data are shown for a given date range. Tableau Resource Monitoring Tool shows up to 80 points on each line chart by default. This prevents lengthy date ranges from drawing too many data points on the screen, which would cause performance problems.

When looking at one hour, each point represents 45 seconds of data. When looking at one full day, each point represents 18 minutes of data. The data is aggregated from one point until the next point. That is to say that given a full day, a point at “10:00” represents an aggregate of all the data from “10:00” to “10:18”.

Line Chart Resolution Figure 1
7 Users aggregated from 10:00 to 10:18
Line Chart Resolution Figure 2
6 Users aggregated from 10:18 to 10:36

Line Chart Date Range

Each point on the line chart represents a span of time (see Line Chart Resolution for more info) and the chart is constructed so that the span is the same length for each point. In order to do this, the exact start date and end date of the chart may need to vary slightly from the user specified date range.

For example when looking at a twelve hour range from “12:10 am” to “12:10 pm”, with an 80 point resolution, each point represents 9 minutes of data. The first point will be adjusted to begin at “12:06 am” and the last point will be adjusted to begin at “12:06 pm”. These times are calculated by finding the first 9 minute interval from 1/1/1970 00:00:00 UTC that contains the specified start date and the first 9 minute interval that contains the specified end date.

Note: research “Unix time” if you are interested in learning more about why the interval calculation begins at 1/1/1970 00:00:00 UTC.

Line Chart Date Range Figure 1
The chart begins with values aggregated from 12:06 to 12:15 am
Line Chart Date Range Figure 2
the chart ends with values aggregated from 12:06 to 12:15 pm

Environment Overview

Dashboard that uses tabs to separate charts into a few categories that are outlined below.

Environment Tab

Environment Tab Overview
Environment Tab Overview

Performance

Performance Chart
Performance Chart

Each point is an average of the data sampled.

Current performance values are sampled using WMI once every minute by default. In multicore environments, every CPU is aggregated for the Processor chart. To see each CPU individually, select the server by it’s name from the Performance dashboard menu.

The Processor, Memory, Disk Queue, and Network charts are all calculated this way.

Network utilization for each server is calculated as follows:

  • The Network percentage shown is the average percentage of bandwidth used over the preceding 60 seconds of time. For example, if within the past 60 seconds, the server had 45 seconds of 0% utilization and 15 seconds of 12% utilization, the average over the past 60 seconds would be 3% and that would be the value you would see for the Network chart.

  • The value of the maximum bandwidth uses WMI(Link opens in a new window). The Resource Monitoring Tool uses the value returned from Win32_PerfFormattedData_Tcpip_NetworkInterface's CurrentBandwidth.

    Note: This value may not always match what you might see reported using other tools. For example, a situation where the virtual adapter reports speed of 10Gbps to Windows, but is running on hardware that only supports 1Gbps. In this situation, because the value returned by the system is 10Gbps, reporting on the Resource Monitoring Tool will use 10Gbps as the maximum speed.

 

Tableau Processes

Tableau Processes
Tableau Processes

Each point is an average of the CPU data sampled for the given Tableau process.

Environments with multiple agents will have a chart for each agent. In this scenario, it’s normal for some agents to have no information about a Tableau process, as some workers could have no processes for a given Tableau process assigned to it.

Background Tasks

Each point is a summation of background tasks that are running or just finished.

For example, a Refresh Extract task will be counted for as long as it’s running, including the point where it finishes. This accounts for background tasks that are very quick. Lengthy date ranges will yield larger tasks counts because this chart uses a summation of activity. See Line Chart Resolution at the top of this article for more information.

Background Task Figure 1
84 total tasks reported at 9:45 when viewing 5 days of data
Background Task Figure 2
4 total tasks reported at 9:45 when viewing 1 hour of the same data

All Tasks is a summation of Refresh Extracts, Subscriptions, and all other background tasks. Navigate to the Background Tasks page to see more detail about each task accounted for.

Concurrent Users

Each point is a summation of the unique Tableau user activity.

Gateway requests like viewing a worksheet or dashboard are activities. If a unique Tableau user does two activities within one point, then it will only be counted once. Lengthy date ranges will yield larger Concurrent User counts because this chart uses a summation of activity. See Line Chart Resolution at the top of this document for more information.

Concurrent Users Figure 1
12 unique Tableau Users reported at 10:30 when viewing 5 days of data
Concurrent Users Figure 2
4 unique Tableau Users reported at 10:30 when viewing 1 hour of the same data

Load Times

View Load Times
View Load Times
View Load Times

Each point of the Longest line represents the maximum duration of all view loads at that point. Each point of the Average line represents the average duration of all view loads at that point. Individual view load durations can be found on the View Loads page.

External Data Requests

External Data Requests
External Data Requests

External data requests are requests to the data server from an IP address that is not one of the machines in the Tableau cluster (not a worker, not the master). Each point of the Longest line represents the maximum duration of all external requests at that point. Each point of the Average line represents the average duration of all external requests at that point.

Background Tasks

Refer to the Background Tasks chart on the Environment Overview page.

Server Tab

Servers Tab The Servers tab has all of the charts that were previously on the Server Dashboard and is focused on monitoring the hardware performance of your servers.

The charts will show aggregated performance data for your entire environment.

Each chart will have a line series per server in the environment.

Refer to the Performance chart on the Environment Tab page.

Insights Tab

Insights Tab
Insights Tab

Slowest Views

Slowest Views
Slowest Views

This chart shows the top five slowest views ordered by average view load duration. The orange dot indicator represents that view’s longest load duration

Longest Extract Refreshes

Longest Extract Refreshes
Longest Extract Refreshes

This chart shows the top five slowest Refresh Extract tasks ordered by duration. Both data source extracts and workbook extracts are considered. They are distinguishable by the icon next to their name.

VizQL Session Environment Activity

VizQL Session Details - VizQL Processor and VizQL Memory spark lines
VizQL Session Details - VizQL Processor and VizQL Memory spark lines

These charts show the same information as the Environment Overview charts. The VizQL Session’s time frame is used as the date range.

View Load Request Details

Historic Load Times

View Load Request Details - Historic Load Times
View Load Request Details - Historic Load Times

This chart shows up to 14 days of load times for the selected view. Each point of the Longest line represents the maximum duration of view loads at that point. Each point of the Average line represents the average duration of view loads at that point.

View Load Environment Activity

These charts show the same information as the Environment Overview charts. Typically, 5 minutes before and after the time the view was loaded is used as the date range.

Who can do this

Any Resource Monitoring Tool user can view the charts.

Thanks for your feedback! There was an error submitting your feedback. Please try again.