If you are on the last page of a set of data and delete the last item from the page the pager does not reset to the previous page. This seems like an obvious bug to me.

Steps to reproduce:

Go to Grid Demo: http://demos.telerik.com/kendo-ui/grid/editing-inline
Select "Edit this example"
Change the pageSize of the dataSource from 20 to 2
Run the demo
Go to the last page
delete the two items on the page

Result: Grid does not move to previous page and incorrectly displays the item number as higher than the total number of items.

We have the grid data-source and the headers bound to the members of the data-source data .
So the grid header on sort pass the data-field to the server .There should be option to change the header option and the data-field .

Cause we might send a VM  of the navigational property to client side and the grid on header click should send original data-field then the VM property

The current method of selecting the item to filter forgeinkey column is not useful for a lot of records.
It is necessary to enable the filter template customization, or add search/autocomplete to current filter template.

KendoUI Grid persist row selection state after update a field programmatically via the set method of the dataItem.

We use the Kendo Grid with a DataSource connected to an OData service. We also use the OOB Kendo Grid functionality to hide columns in the UI (column.hidden = true). Through testing we found out that the Kendo grid does not use the OData option of selecting specific columns for retrieval. The grid will request all columns for a table, which leads to overhead in the amount of data sent from the server to the client. We sometimes have tables with 50 columns, for which only 10 columns are shown.

When grid column is hidden template code is still being executed. For large data sets with many hidden columns this makes grid scroll slower than is necessary...
{ id: 'Artist', title: Columns.ARTIST, field: Columns.ARTIST, width: '100px', hidden: true,
	template: function (dataItem) //this gets executed even when column is hidden. 
	{
		return someMethodToDetermineArtrist(dataItem);
	}
}

If Column widths are not set, The grid column and column header will be misaligned when autoscroll is enabled unless the following code is used which feels like a hack. This code snippet was given to me by a support team member.

    dataBound: function(e)
    {
      
        if($(".k-grid-content table").height() > $(".k-grid-content").height())
        {
          	$("#grid1 div.k-grid-header").css('padding-right',kendo.support.scrollbar());

            $("#grid1 div.k-grid-content").css('overflow-y', 'visible');  
        }
       	else
        {
            $("#grid1 div.k-grid-header").css('padding-right', '0');

            $("#grid1 div.k-grid-content").css('overflow-y', 'hidden');        
        }
    }

When using a template for Grid PDF export, usage of AngularJS scope variables should be possible, e.g. to customize the title.

I believe drag-drop, resize and reorder functionalities should be accessible using keyboard for people with very poor eyesight.

Grid sorting needs to sort Enum by text, not by their numeric undelaying value. Or at least it should be an option what to sort by.

When a grid is first showed with a group value (e.g. a date value), it should be possible to apply a sorting direction to the grouped value (e.g. descending)

Expose a clean way of destroying kendoDraggable.

For instance at the moment I only want selected rows in a grid to be draggable, to when a row is selected I make it draggable, but then I need to clear all other draggables in the grid.  Can't find a clean way to do this.

In lieu or as part of select-all capability:

Expose a way to retrieve the filtered set of data (or keys) in the grid and other filterable controls.

In other words, when a user has applied their own filters, there is currently no method to retrieve the filtered set across the entire dataset. That makes implementing a select-all button a bit tricky as you have to either select all on the current page (view) or the entire unfiltered dataset and then apply the filter yourself.

With the kendoGrid, if no columns are specified, all columns are automatically rendered. However, dates (at least, I have not tested other non-strings) are rendered in non-human-readable form (at least if they are from SQL Server).

Since date rendering is tied to culture parameters, it would make sense to specify culture at the grid object level and have dates, currency, etc. render according to that. It would also be great to have filters setup for these different types applied to dynamic columns.

The Kendo grid is slow to render row details when using Internet Explorer.   The reason, according to a forum post, is because IE does not handle the ":not()" selector well, which the Kendo uses.  Fix this problem.

It would be great to have a couple of CSS classes to do customization of the style of scrollbars in Kendo UI Web controls (multi-browser) because otherwise its really complicated to integrate javascript plugins to do this in the Kendo UI structure.

I want to be able to change whether a grid is scrollable, as its datasource grows or shrinks, without having to destroy and recreate it, and I want to control whether the columns are equal width or auto-sized, independently of whether it's scrollable. I also want to change its height, and turn on and off the option that lets it auto-size itself.

(Also I'm running into a similar issue with charts, where I can't change the tooltip template after it's set.)

Can we please have a way to save and retrieve the raw html formatted as we'd like.

Currently it does not matter how you write the raw html, it will come back in an unusable blob of html

When the user types a value in a grid column filter, clicks the filter button, then deletes the value from the column filter, and clicks the filter button, the grid results will not change to reflect the updated filter. When the filter text is removed, and the filter button is clicked, the grid results should be updated.

By default Kendo Grid places a new created item/row at top of the grid in batch / incell edit mode. When saved the item remains top. The items count is increased in this case.
A click on Cancel puts the new item to the end of the grid and the items count is decreased.
There is just a work-around which stops the propagation of the cancel event when the datasource has no chances (thanks to Petur). 
That should be the standard behaviour.