I suggest you ...

Add ability to view Azure Table size/entity count (rows)

Hi Mike,

I've created this idea as suggested on the forums: http://social.msdn.microsoft.com/Forums/en/windowsazure/thread/ea18ae29-36a3-42c6-8420-877216efbd42

One of the big challenges in adopting azure table storage over traditional SQL storage is the ability to know how much data is stored and how it is being used.

Being able to break size/rows down by partition would be invaluable when trying to modify / optimize Partition & Row keys. (Given data doesn't always grow as we would expect, and new bottlenecks can & will emerge).

In addition the ability to view the usage data per table / partition would be fantastic.

Obviously there are ways of doing this at a higher layer, but having a consistent and reliable API for exploring the data would go a long way to making table storage a viable NOSQL alternative.

Thanks
James

66 votes
Vote 0 votes Vote Vote
Vote
Sign in
Check!
(thinking…)
Reset
or sign in with
  • facebook
  • google
    Password icon
    I agree to the terms of service

    You'll receive a confirmation email with a link to create a password (optional).

    Signed in as (Sign out)
    You have left! (?) (thinking…)
    James AndrewsJames Andrews shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

    2 comments

    Sign in
    Check!
    (thinking…)
    Reset
    or sign in with
    • facebook
    • google
      Password icon
      I agree to the terms of service

      You'll receive a confirmation email with a link to create a password (optional).

      Signed in as (Sign out)
      Submitting...
      • James AndrewsJames Andrews commented  ·   ·  Flag as inappropriate

        Hi Mike,

        Thankyou - the level of engagement & transparency of the Azure team is one of the most appealing features of the platform.

        Below are some additional points for the storage team.

        This feature would not need to function as a SELECT COUNT(*) equivalent, ie: the results don't need to be realtime or 100% accurate. Similar to SQL Server when using the sysobjects table for row counts, and sp_spaceused for sizing.

        I would think this provides a solution to several problems:

        1. Sub billing of the platform (As opposed to creating an entire billing solution as suggested elsewhere in the forum, if a PK used a customer ID this would solve allow for end-user based usage breakdowns)
        2. Optimization of partition / row key schemes (& hence query performance)
        3. An understanding of where resources are being consumed (when I look at my bill I might get a shock, it would be good to know exactly what was taking up space)

        I thought it might be interesting to provide this data through a read only table, similar to SQL Server. (Although I don't really mind, as the first thing I would do is create a table to store a history of table / partition size & count so I could monitor growth over time... but others might find this a nice way of accessing the data)

        Thanks
        James

      Knowledge Base and Helpdesk