It would be nice to be able to specify in a datasource filter two fields and an operator for the sake of data reporting. For example, where "completed date" is greater than "promised date" or "Actual cost" is greater than "estimated cost."

This sort of thing can be accomplished by calculated fields; a filter of where "cost overage" is greater than 0. The down side of the calculated field approach is all comparisons need to be pre-defined. 

At the moment, the DataSource only uses one field for its ID property (if you enter several, then only one is used, apparently).  This doesn't reflect reality, in that in many cases (where one record is a child of another), the key field for a database table is a composite (from 2 or more fields).  
The existing situation causes problems in inline editing of the KendoUI grid, because default values need to be set for foreign key fields, but if these are set to a valid value, then spurious calls to the Inline_Create method occur because the DataSource treats the record as a new record when it is not.
In any case, allowing multiple fields would simply reflect reality.

Provide a data source setting and functionality that all the dates be created as UTC replacing the need to write the requestEnd code everywhere:

https://docs.telerik.com/aspnet-mvc/html-helpers/data-management/grid/how-to/editing/utc-time-on-both-server-and-client

It would be very helpful not to have to do this in every scenario that requires using UTC dates and have the data source create the dates in UTC as they reach the client.

When the grid is configured for working with UTC on the server-side, the value sent from the filter menu on the client-side would not be properly sent to the server. 

If the stringifyDates option of the transport of the data source is set, the value would still be sent in a localized date format without information for the time zone, for instance, GMT(+0300).

In case the value sent from the client is modified, the parameterMap would be called afterward and would override the modified value. 

Moreover, the Kendo UI Binder on the server-side is not configured to be working with dates that have additional data for the time zone appended to the end. Therefore, sending the modified data would not be parsed correctly and FormatException error would be thrown.

Ignore language specific characters when filtering. Example: search phrase "Deja" should find word "Déjà".

Trying to bind to the 'id' and 'for' properties gives an error that this is not supoprted. Both are basic bindings and should be suported.

We use the new version of SignalR which is 2.2.3 currently and it seems the Kendo DataSource cannot bind to it. It would be awesome if Kendo supported it. 
Here is a link: https://www.nuget.org/packages/Microsoft.AspNet.SignalR.Core/

I would like to be able to dynamically update the text in my button by binding to a viewmodel.

Thanks

We need an option to make the datasource immediately drop the current in-flight data fetch request if another one comes in.

Currently, if the user clicks around several times in a row changing grid datasource parameters too quickly, several requests get fired off, but the grid waits for the very first request to complete before issuing the next one. This makes the user's wait time twice as long and also makes the grid content appear to switch back and forth as requests complete one after another. Very unpleasant effect.

Right aligned columns are common for number fields and it would be nice to simply set a property in HTML and/or the view-model:
E.g { "field": "amount", "align": "right", "title": "Amount", "width": "110" }

Currently when the add() method is called on a DataSource, if the id, as defined in the Model, is included, then the sync method is called, the data is not persisted to the _pristineData object, and neither the Create nor Update transports are called.  This has the effect of not persisting the data and so if the cancel button is pushed on an edit screen, the added object is mysteriously dropped.  

Please see my post on the kendo forums for a more detailed explanation including demos of the issue.  http://www.telerik.com/forums/if-id-is-provided-in-add()-sync-does-not-update-_pristinedata-and-create-update-are-not-invoked

I would suggest that anytime the add() method is used, it should be considered a 'new' record, or at least when the id provided does not match with any other id currently in the data array. If it matches a record in the data array, run the Update transport, otherwise, run the Create transport.  Simply including a value for the id should not remove the record from ever being persisted. 

This feature was already requested in another post, but it was missunderstood. (http://kendoui-feedback.telerik.com/forums/127393-kendo-ui-feedback/suggestions/11578722-support-sorting-datasource-groups)

The idea is to allow for Groups in DataSources to be sorted in ways other than alphabetically; this will allow for certain groups to be given priority.

When using aggregates, it should be possible to define your own aggregate function, or even set it globally for kendo, so it could be used in other places. i.e. 

currently you can only specify the function name like 'sum', 'avg', etc. with this idea you could give the whole function like function(arguments...) {}

Also from what I saw it would be nice to have the `functions` object from kendo.data.js accessible, so you could add your custom functions, like:
`kendo.data.aggregate_functions['median'] = function(arguments...) {}` 

This is more of a feature request...

If I want to refresh a datasource every 60 seconds, be it a grid, a dropdown, etc., I have to add a snippet of code something like this:

var interval = 60000;

         setInterval(function()

         {
             $("#my-grid").data("kendoGrid").dataSource.read();
         }, interval);

Yes, I realize this is trivial, but it would be nice if you could add an AutoRefresh function directly to the DataSource API that handled all of this under the covers.  Example:

http://odata.github.io/WebApi/#04-26-InOperator

OData v4 now supports the IN operator as a short hand for multiple or queries (as of http://docs.oasis-open.org/odata/new-in-odata/v4.01/cn01/new-in-odata-v4.01-cn01.html#_Toc485385090 ). This could help to change code such as:

var selectedStations = $("#cmbStations").data("kendoMultiSelect").dataItems();
            var stationIds = [];
            for (var i = 0; i < selectedStations.length; i++) {
                stationIds.push({
                    field: "StationId",
                    operator: "eq",
                    value: stationId
                });
            }
            dataSource.filter({
                logic: "or",
                filters: stationIds
            });

into a much cleaner:

var selectedStations = $("#cmbStations").data("kendoMultiSelect").dataItems();
            var stationIds = [];
            for (var i = 0; i < selectedStations.length; i++) {
                stationIds.push(selectedStations[i].StationId);
            }
            dataSource.filter({
                field: "StationId",
                operator: "in",
                value: stationId
            });

The DataSource schema must be given a total value in order to page correctly. However, DataSource expects this value to come from a response body field. In some cases, we must work with systems that return the total value as a custom field in the response header. We need a way to read custom response header values. 

Yes, the DataSource transport uses jQuery.ajax, so we can implement a custom complete function to read the response headers from the jqXHR. However, complete fires AFTER schema.total is set so it does no good.

For easy using local arrays with kendo DataSource it would be good to have a functionallity to assign an array (kendo.obersvableHierarchy) without configure transport for CRUD operations.
When using Kendo MVVM it's easy then to define a grid the columns and field, and bind only the observable array as datasource.

And it would be great to have the possibility to assing depended arrays as field in an array that will be used as datasource for a grid or dropdown. Then we can assing the depended array of a selected dataitem to another control. So there is an need to have a special datatype like "dependedArray" and changes to this depended array has no effect to all ohter controls that where the viewmodel was bound.
Then we have more time to develope business logic. 

One of key areas for enterprises is data security. If we can add the capability of encrypting data via kendo datasource, it would be awesome. 
This encryption should automatically happen when data is received from a service using the encryption algorithm specified as an attribute of kendo datasource. Similarly, decryption should also happen when the datasource receives the encrypted data.

See this forum post for details: http://www.kendoui.com/forums/framework/mvvm/getting-access-to-array-index-while-binding-an-array-source.aspx