Recent AWS Customer Success Stories & Videos

More AWS Customer Success Stories...

« New Features for Amazon CloudFront: Server Name Indication (SNI) and HTTP Redirection | Main | Access Logs for Elastic Load Balancers »


Feed You can follow this conversation by subscribing to the comment feed for this post.


Great - I've tried to set this up "manually" but failed. I was surprised that a cluster was started for each table tho. A bit of overkill for my 10 item table!

Andrey Chursin

This is great and very expected feature. However I am curious about some implementation details.
When we tried to make our own DDB backup we stacked with issue, that current DDB API with scan/get operations is insufficient to make consistent backup of database state.
In our application we are using consistent reads and CAS a lot and our logic really rely on ordering of such things.

I mean, imagine our application updates entity A, then updates entity B. Then there are tree 'consistent' states of database -
1) A and B aren't updated
2) A is updated but B is not updated, and
3) both A and B are updated.

However, if you are making simple DDB API calls like scan/get during backup you can have snapshot of DDB state when entity B is updated and A is not, which is illegal backup state, as far as application can not reach this state in 'normal' way.

So my question is - do you make some internal API calls and ensure that writes to DDB after some point is rejected and backup is consistent.
This practically means - can I run backup on 'live' application, or do I need to stop it so nothing updates data during backup(which practically means I can not use backup)


Thank you! This was sorely needed.

I must say that aws pipelines are extremely confusing. So, I'm glad you're hiding the complexity with this new feature.

Simon McLean

we have exactly the same question as Andrey Chursin - is the backup a consistent snapshot of a table?


I wish you guys would make your pricing easier to calculate.

At the bottom of the dynamodb article, it says you will be charged for using this one feature by three different services (S3, Pipeline, and EMR). It's very hard to justify these costs when you don't even know what they'll be.


That's a great feature which was needed!!! I started to use it and it gave my company a great value and easy to configure for cases of backup & recovery. I didn't found the option in case that we would like to take DB backups more than one time a day, is it possible? (I just configured it yesterday and was tired).

There were some companies that tried to sell it but it was not trustable and with one I faced a bug which took us lot of time to fix it, we lost valuable data, despite this company was on AWS market place, it was an unknown company which when the product was really needed it only cause us damage.

We are about to upload part of our company on Amazon and this will accelerate the process.

Good work AWS!!! Thanks.


Now I see that via the AWS Management Console you can schedule to backup at any time and as many times as you want per day!

Amazing product, since I'm a simple employee in the mentioned company, I'll push this and hope to get some credit/promotion.

Please don't publish the company name inserted in the URL.

Thanks again.

The comments to this entry are closed.

Featured Events

The AWS Report

Brought to You By

Jeff Barr (@jeffbarr):

Jinesh Varia (@jinman):

Email Subscription

Enter your email address:

Delivered by FeedBurner

April 2014

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30