I save the diagram json to everlive by using the save method and load it into the widget with the load method.  My problem is that connection blocks cannot be edited now.  I don't think CRUD should be required for this ability.  

I find the documentation is critically lacking with regards to the available options for a feature.

Here is an example of what I mean...

http://docs.telerik.com/kendo-ui/api/javascript/dataviz/diagram/layout#configuration-alignContent

The diagram Layout > alignContent feature "Specifies the alignment of the content."

Ok, but the documentation does not even say what the allowed values are. Instead, it just has "alignContent String (default: "start")".

Now, I guess the allowed values here are the same values allowed for an SVG text-anchor.
https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/text-anchor

How about adding those possible values to the api documentation, or including a link to the relevant area of the SVG specification?

Hope this makes sense.

You should include a flowchart demo for your diagram tool. It is possible - somewhat - but you are not even including a demo for it, while vendors like syncfusion do! 

The way I found out a flowchart was possible was because of the Kendo UI cookbook and cookbook code files. 
https://www.packtpub.com/web-development/kendo-ui-cookbook

Seriously, I don't know why you are not including a demo for a flowchart in your demo list for the diagram tool; this should be a no-brainer to include, and in theory, you should not even have to build anything new, just use the diagram tool you have already made. 

It would love to have an export feature for Diagram which will give the possibilty to make an export to Microsoft Visio.

In large Diagrams the connections will route trough the shapes itself what is very confusing. I would like to have an option to avoid this behaviour.

Rad Diagram - Support copy paste from one diagram to another,

Expand/Collapse feature on click of the diagram nodes. 
Load data on demand during enpansion of the node.

I want the content to be divide into two lines. So I tried setting `connectionDefaults.content.visual` with the below function:


            var getConnection = function (data) {
            var g = new kendo.dataviz.diagram.Group({
                autoSize: true
            });
            var text1 = new kendo.dataviz.diagram.TextBlock({
                text: data.label.name,
                fontSize: 16
            });


            g.append(text1);

            var text2 = new kendo.dataviz.diagram.TextBlock({
                text: data.label.value,
                fontSize: 16
            });


            g.append(text2);
            return g;
        }
But the problem here is that the `data` field does not populate the connection data due to which assigning values is not possible.


A solution for this is much appreciated.

When connections are detached from shapes or drawn via API with coords instead of shape ends, they don't move with selection when multiple shapes/connections are selected.  When these "floating" connections are selected in a multiselect scenario, they should move with the rest of the selection.

Currently, a shape that is a group of inner shapes is selected by clicking anywhere inside its bounding box. 
However, when there's a gap between the inner shapes through which you see another shapes, clicking that other shape will not select it due to the bounding box.
If instead the selection was based on the inner shapes, that wouldn't be a problem.

as demo'd here : https://gojs.net/latest/samples/sankey.html

Its good to have minimize and maximize the shapes drawn with kendo diagram as seen in this example
https://gojs.net/latest/samples/IVRtree.html

it will very useful when we are drawing with a huge data. we can clearly see the actual flow as required

It would be nice if the diagram control could have undo/redo event.

It would be nice if the diagram control could support undobale properties load back from saved data. And also if shapes, connections could acquire this property (not only through diagram.addShape and diagram.addConnection).

It would be nice if the diagram control could support complex shape json export (diagram.save()) and import (diagram.load()), like shapes contains TextBlocks, etc.

Add built-in support for guidelines that will be visible as an infinite grid of lines, with a snap-to capability.

What about adding dynamic scrollbars to the HTML5 Diagram widget?

As developer I'm familiar with the pan and zoom concept, but customers and normal users are expected to see scroll bars to move around a diagram. 
In this demo

http://demos.telerik.com/kendo-ui/html5-diagram-sample-app

you can add objects that fall outside of the main view, but no scroll bars appears.

Here http://kendoui-feedback.telerik.com/forums/127393-telerik-kendo-ui-feedback/suggestions/6552432-html5-diagram-control-allow-scrolling

was posted a similar request, but the scrollbars Kenshin suggested 
are static.

The scrollbars does't readapt if the work area changes. 

If you move objects around, apply a different zoom, add more elements, and so on.

It would be nice if, in addition to "polyline" and "cascading", if there were other options for the shape of the diagram connectors.  I would like to see a "spline" option that possibly follows a similar path to the "cascading" square shape, but flows with curves.   Maybe even a configuration object, instead of a string that could define the parameters of the shape.

{ 
type: "spline", 
mode: "bezier", // or "natual" or any other spline algorithm //
maxCurves: 3, // max number of inversion points in the spline //
curveRadius: 100 // size in pixels //
// any additional properties for configuring the curve, like bezier length, instead of curve radius &c. //
}

The diagram has its own built in flow layouts (left, down, &c.).  I would like to see an interface provided to allow a custom callback to allow a programmer to create their own layout flow designs.  All positioning calculations would be up to the user implementation, but objects that provide information about the current layout while rendering could be provided as parameters to the callback.

At the moment, "resizing" shapes is merely a scale transformation. 
Proper resizing would be much better looking!