Skip to end of banner
Go to start of banner

Embedding Dashboards and Charts

Skip to end of metadata
Go to start of metadata

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

« Previous Version 28 Next »

You can find a live code example that details how to embed the dashboard in your dashboard installation.

If your dashboard is installed locally with all the default settings, to access the example go to http://localhost:8224/pi_embed_example/.

Embedding A Panintelligence Chart

Step 1 - Ensure your dashboard is running in HTTPS

See Configuring HTTPS for more information.

Step 2 - Configuring the Dashboard

By default, embedding the dashboard in an iframe will no longer work without configuration. This is by design to provide the most secure out of the box configuration.

We have implemented click-jacking protection in the form of the recommended 'frame-ancestors'. By default, this is configured to disallow any embedding from any site. This can be changed via the Panintelligence Configuration Tool.

Using the Configuration Tool

On Windows

You can get to the Configuration Tool via the Start menu:

Or by navigating to the Dashboard installation folder. On Windows, the default location is C:\Program Files\Dashboard\Dashboard:

Please ensure that you run configuration-tool-GUI.exe as an administrator.

On Linux

Double click configuration-tool-GUI.AppImage.

Editing the CORS / Frame Ancestors / Trusted Hosts and the Cookie settings

If you need embedding to work with Internet Explorer, you must set X Frame Host option too. Keep in mind that we do not maintain support for Internet Explorer.

From the Trusted Hosts dropdown you’ll see the following values:

  • none: No embedding allowed

  • self: Embedding allowed from the same host

  • custom: Embedding allowed from the hosts specified in the space separated list

Protocol (http/https) is required for trusted hosts / frame ancestors when using a wildcard.

Using environment variables

Settings can be applied via environment variables and this is true for these fields too:

PI_TOMCAT_FRAME_ANCESTORS="http://localhost:5000 https://my.awesome.app"
PI_TOMCAT_COOKIE_SECURE=true
PI_TOMCAT_COOKIE_SAMESITE=none

If you need embedding to work with Internet Explorer, you must set PI_TOMCAT_X_FRAME_OPTIONS option too. Keep in mind that we do not maintain support for Internet Explorer.

Protocol (http/https) is required for trusted hosts / frame ancestors when using a wildcard.

Using the configuration file

These settings can also be applied via a configuration file to support unattended installs and GUI-less environments.

To do this, find or create a file named dashboard.json next to the configuration tool:

NOTE: When using the configuration tool by the command line and you’re specifying a configuration file, you must specify an absolute path to it (not a relative path).

  • ✔ Absolute path: configuration-tool configure --json-file=/absolute/path/here.json

  • ✘ Relative path configuration-tool configure --json-file=relative/path/here.json

Then edit it using your favourite editor. You’ll need to change tomcat.cookies.secure, tomcat.cookies.sameSite and security.trustedHosts.

If you need embedding to work with Internet Explorer, you must set xFrameOptions option too. Keep in mind that we do not maintain support for Internet Explorer.

Protocol (http/https) is required for trusted hosts / frame ancestors when using a wildcard.

Step 3 - Getting a token

You can run a working example by following this repository:

In our examples, we’re doing this in the frontend just so you can follow a full example.

However do not do this in production, you’ll be exposing your password!

Typically your backend communicates with the dashboard in order to get the token. Once in place, your frontend then just uses it.

You can use our API to fetch a token for any user. Below we’ve got an example in javascript using jquery:

var url = "https://localhost:8224/" //Change this to your url
var username = "admin" //Change this to the user you want to login as
var password = "dashboard" //Change this to user's password
$.ajax({
    // POST to the API with username and password to swap that for a session token
    type: "POST",
    url: url + apiToken,
    contentType: "application/json",
    xhrFields: {
        withCredentials: true
    },
    beforeSend: function (xhr, settings) {
        xhr.setRequestHeader("Authorization", "basic " + username + ":" + password);
    },
    success: function (data) {
        login(data.token);
    }
});

function login(token) {
    // Issue an ajax request to embeddedTokenLogin with the token in the header to pick up the session cookies
    $.ajax({
        type: "GET",
        url: url + "pi/auth/embeddedTokenLogin",
        contentType: "text/HTML",
        xhrFields: {
            withCredentials: true
        },
        beforeSend: function (xhr, settings) {
            // This is the header we need to set
            xhr.setRequestHeader("Authorization", "bearer " + token);
        },
        success: function (data) {
            // Once we have a succesful ajax request with authentication we can load the charts in the iframes
        }
    });
}

Step 4 - Embedding

Chart

To embed a chart you’ll need an iframe and the chart’s ID. You can get this by going into the dashboard, clicking chart tools at the top-right of a chart cell and selecting the i icon:

Then copy the code snippet:

 

<iframe id="chart_iframe" src='https://your.dashboard.com/pi/chart#chart-filter/481__1' width="500" height="500"></iframe>

You should probably only update/show the iframes once the user has been logged in.

That would change the login code to be more like:

function login(token) {
    // Issue an ajax request to embeddedTokenLogin with the token in the header to pick up the session cookies
    $.ajax({
        type: "GET",
        url: url + "pi/auth/embeddedTokenLogin",
        contentType: "text/HTML",
        xhrFields: {
            withCredentials: true
        },
        beforeSend: function (xhr, settings) {
            // This is the header we need to set
            xhr.setRequestHeader("Authorization", "bearer " + token);
        },
        success: function (data) {
            document.querySelector("#chart_iframe").src = "https://your.dashboard.com/pi/chart#chart-filter/481__1"
        }
    });
}

Limiting drill levels on embedded charts

When embedding a chart that contains multiple drill levels, it is possible to use the query parameters “drillStart” and “drillEnd” in the URL to limit the drill levels that a user can see.

When using drillStart, the URL will also need to be configured to include the chart ID and hierarchy position of each chart that is being skipped and the chart at the drillStart level. These must be separated by the !$! delimiter when adding multiple levels. When drilling to the top, the user will return the the level defined as the drillStart parameter.

It is possible to use the parameters individually or together. The parameters are 0 indexed - with the first hierarchy in the chart corresponding to drill level 0. The functionality will also work for dynamic drill conditions.

The below url would allow the user to see drill levels 1, 2 and 3 (i.e the 2nd, 3rd and 4th hierarchy levels) of the chart:

http://localhost:1234/pi/chart/?drillStart=1&drillEnd=3#chart-filter/251__1/d-f-p=251__0/!$!251__1

*Location and syntax of the query parameters.

*Example of identifiers added to view level 1 (2nd hierarchy level ) of the given chart.

Category

We have embedded a chart but what if we want to embed a category? Embedding a category isn’t actually any different from embedding the whole dashboard (yes, you can do that too!).

Below are the different category modes that you can use when embedding the categories:

categoryMode=0: Normal

This will show the category exactly as it appears on the dashboard.

categoryMode=1: Headless Single Category

This will hide the category header and the category list.

categoryMode=2: Single Category

This will show the category header options but will still hide the category list.

categoryMode=3 Headless Normal

This will remove the header but keep the category list.

 

Interacting with the left & right drawers

One thing to note with the different embed options is that you can actually use JavaScript’s post message API to open the category list / filter panel within the various modes if you wish to (and if they’re available in that category mode).

The left drawer tag will open the category list as long as the embed mode allows for the category list to be active (Options 0 and 3). To interact with it you can use something like below:

targetWindow.postMessage("open-drawer:left");

The right drawer tag will open the filter panel for every embed mode available. Similarly, you can open it using the same syntax as above:

targetWindow.postMessage("open-drawer:right");

The right drawer panel can be opened via a query parameter too: enableDynamicFilterPanel=true. This lets the dashboard know that the category should load with the Category Object Filters list open

The full range of messages the dashboard interprets can be found in our API Embed Examples and they include opening, closing, pinning and unpinning drawers.

Using filters to filter a chart via JavaScript while embedded

With Category Filters (whole category)

Adding filters

The dashboard supports embedding with default category filters selected.

Refer to Embedded Categories for details. If changing the filter values outside of the dashboard is needed, we will need to implement functions to be called outside of an iframe, so that filter values are supplied to the dashboard elegantly.

 

The current structure of the URL for category filters

IMPORTANT: This is just for information. We don’t recommend editing the URL directly to change filter values, because it’s not a reliable approach and the implementation may break when upgrading the dashboard

 

Category-level filters are applied to all the charts in the category, it currently has such a pattern pattern:

scheme://host:port/pi/#PA-CA###category/id__level/c-o-p=FILTERS

For example:

This category has filters Department = Dev:

https://localhost:8224/pi/#PA-CA###category/141__0/c-o-p=eq__[[Department]]{{dev}}

And this one has Department = Dev and Number > 100:

https://localhost:8224/pi/#PA-CA###category/141__0/c-o-p=eq__[[Department]]{{dev}}/*$*gt__[[Number]]{{100}}

With Temporary Filters (per chart)

Current URL Structure

IMPORTANT: This is just for information. We don’t recommend editing the URL directly to change filter values, because it’s not a reliable approach and the implementation may break when upgrading the dashboard

Temporary filters have something with the following pattern:

scheme://host:port/pi/chart#chart-filter/id/t-f-p=FILTERS

For example:

Filtering the chart with Department = Dev:

https://localhost:8224/pi/chart#chart-filter/330__1/t-f-p=eq__[[Department]]{{dev}}

Filtering the chart with Department = Dev and Number > 100:

https://localhost:8224/pi/chart#chart-filter/330__1/t-f-p=eq__[[Department]]{{dev}}/*$*gt__[[Number]]{{100}}

Types of filters

IMPORTANT: This is just for information. We don’t recommend editing the URL directly to change filter values, because it’s not a reliable approach and the implementation may break when upgrading the dashboard

Filter Name

Pattern

Object Types

Examples

Equals

eq__[[Object]]{{Value}}

Dimension

eq__[[Region]]{{North}}

Measure

eq__[[Count]]{{50}}

Greater than

gt__[[Object]]{{Value}}

Dimension

gt__[[Region]]{{North}}

Measure

gt__[[Count]]{{50}}

Lower than

lt__[[Object]]{{Value}}

Dimension

lt__[[Region]]{{North}}

Measure

lt__[[Count]]{{50}}

Greater or equals

ge__[[Object]]{{Value}}

Dimension

ge__[[Region]]{{North}}

Measure

ge__[[Count]]{{50}}

Lower or equals

le__[[Object]]{{Value}}

Dimension

le__[[Region]]{{North}}

Measure

le__[[Count]]{{50}}

Not equals

ne__[[Object]]{{Value}}

Dimension

ne__[[Region]]{{North}}

Measure

ne__[[Count]]{{50}}

Like

like__[[Region]]{{Value}}

Dimension

like__[[Region]]{{Nor%}}

like__[[Object]]{{'Value'}}

Measure*

like__[[Count]]{{'1%'}}

Not Like

like__[[Region]]{{Value}}

Dimension

nlike__[[Region]]{{Nor%}}

like__[[Object]]{{'Value'}}

Measure*

nlike__[[Count]]{{'1%'}}

Top

top__[[Object]]{{Value}}

Dimension

top__[[Region]]{{5}}

Measure

top__[[Count]]{{5}}

Bottom

btm__[[Object]]{{Value}}

Dimension

btm__[[Region]]{{3}}

Measure

btm__[[Count]]{{3}}

In

in__[[Object]]{{Val1::::Val2::::Val3}}

Dimension

in__[[Region]]{{North::::South::::West}}

Not In

nin__[[Object]]{{Val1::::Val2::::Val3}}

Dimension

nin__[[Region]]{{North::::South::::West}}

Between

btw__[[Object]]{{Value1::::Value2}}

Measure

btw__[[Count]]{{300::::500}}

Not Between

nbtw__[[Object]]{{Value1::::Value2}}

Measure

nbtw__[[Count]]{{300::::500}}

*May not work for some databases

Upper/lower case values make a difference!

Delimiters

/*$* - Used to separate multiple criteria items

:::: - Used to separate values within {{}}, used for Between/Not Between and In/Not In list criteria

Gotchas

CORS errors?

If you see an error in the Javascript Console such as:

Access to XMLHttpRequest at 'https://dashboard-host/' from origin 'https://your-app-host/' has been blocked by CORS policy: No 'Access-Control-Allow-Origin' header is present on the requested resource.

Then you need to review your frame-ancestors settings (via the config tool or via the config file).

If everything still looks correct, then make sure you’re not using any browser addon that attempts to bypass CORS.

Login screen keeps showing / looping after logging in?

Check your cookie settings

This usually means your cookie settings aren’t set up correctly. Review your cookie settings (via the config tool or via the config file).

Is your dashboard behind a proxy?

If the cookie settings look correct, it might be that your dashboard is being served via http and the proxy in front of it will be doing the ssl communication. If that's the case, you'll need to tell the dashboard that it's behind a proxy and that the proxy is doing https. This is done by editing your proxy settings in one of three ways.

Using the configuration tool

Editing the configuration file

Edit dashboard.json configuration file:

{
        "proxy": {
            "enabled": true,
            "scheme": "https",
            "secure": true,
            "host": "proxy-host.com",
            "port": 443
        }
}

Using environment variables

PI_PROXY_ENABLED=true
PI_PROXY_SCHEME=https
PI_PROXY_IS_SECURE=true
PI_PROXY_HOST="proxy-host.com"
PI_PROXY_PORT=443

Is your dashboard a try-pi or an AMI?

This is the same scenario as if having the dashboard behind a proxy.

In settings.sh or in startup.sh, set the following environment variables:

export PI_PROXY_ENABLED=true
export PI_PROXY_SCHEME=https
export PI_PROXY_IS_SECURE=true
export PI_PROXY_HOST="<domain url>"
export PI_PROXY_PORT=443
export PI_TOMCAT_COOKIE_SECURE= true
export PI_TOMCAT_FRAME_ANCESTORS=<URL of client site>
export PI_TOMCAT_COOKIE_SAMESITE=none

Then make sure you reboot your instance.

Check the URL being embedded

We’ve noticed the cookie doesn’t get recognised if there’s a double-slash in the URL of the iframe.

For example: https://dashboard//pi/ or https://dashboard/pi// instead of https://dashboard/pi/

Embedded Sub Domains

If you have subdomains, the hostname in the custom trusted hosts will need to be entered as *.hostname *.hostname 
 
 

  • No labels