Comments on supportable number of users?

Dec 5, 2014 at 5:00 PM
Edited Dec 8, 2014 at 10:31 AM

This is my first potentially large scale use of Azure Tables and I've just had a look at the partitioning strategy for the index table.

We are planning on only allowing registration via external identity providers: Google, FB, Yahoo and MS.

It seems the implemented strategy will only allow a small number of partitions for this table, which could adversely affect performance. But perhaps I've missed something in the implementation?

Would anyone be able to comment on the sort of numbers of users this implementation can support? I'd be particularly interested in hearing from anyone who has worked with large Azure tables in the past.

Dec 8, 2014 at 2:50 PM
Having done some further research, Troy Hunt ( has a great article about creating and querying a large table of email addresses. He used a partition key of the domain and the row key was the user name part of the address e.g.

pk =; rk = foo;

Using this approach his largest partition was 32M+ rows, and performance was still excellent when hitting that partition.

In the comments at the end of the article, someone (JD Conley) has suggested using a hash of the email address to evenly spread the rows across a pre-defined number of partitions. That could be a useful approach for very large tables as well.

Food for thought.
Dec 8, 2014 at 7:36 PM
@chalkie - this is some good feedback. I will add this as a work item for the 2.0 release to evaluate your scalability concerns and probably make an adjustment to the key structure. -DM
Dec 8, 2014 at 7:40 PM
This discussion has been copied to a work item. Click here to go to the work item and continue the discussion.
Sep 6, 2016 at 4:30 PM
Please add this issue to github for followup . Thanks!
Marked as answer by dotnetdavy on 9/6/2016 at 9:30 AM