Please provide the ability to manage backup and restore on our own.
Restore through support is not clear and inconvenient.
It turned out restore is only available through paid support plan.
This document https://docs.microsoft.com/en-us/azure/cosmos-db/online-backup-and-restore#restoring-a-database-from-an-online-backup does not say it in any word.
The situation made our customer extremely unhappy.729 votes
This feature is now available in Preview.
You can read more here.
In a partitioned collection, it would be useful to have the option delete all data with a given partitionId.674 votes
We’re looking to get some feedback on this feature and would love for you to try out a preview.
Really need to be able to set resource level access control integrated with Azure Active Directory. The current built-in user / resource access control is a pain to use and we end up with just using the master key and giving everyone access to everything.
Azure SQL DB already has this, and is a pleasure to work with.
Perhaps you should take a leaf from your sister team on this -377 votes
we are currently using Materialized Views with Cassandra, and we would like this feature to be available with Cosmos DB.
We are currently working on this feature and we would like to get feedback from users here.
Please provide feedback to this form here.
It is buried in the FAQ;s on the pricing page. The current billing structure will take the largest RU offer in a given hour increment and that's what is charged for that hour. If you scale up to 1000RU's at 10:59 and down to 400 at 11:01, the billing for both hours will be @ 1000RU's. This makes scaling up and down harder to deal with and reason about. Can you evaluate changing the billing granularity down to the minute level in order to allow shorter bursts of scale up / down billed for the actual usage.42 votes
We’ve recently announced a serverless model for Azure Cosmos DB – which we believe addresses use-cases with sparse but spiky throughput. Serverless enables billing based on a per consumed request unit basis – which goes 1 step further than managing provisioned throughput at a per minute granularity.
We will continue to support the provisioned throughput model (along with autoscale) for use-cases that have high throughput utilization and/or require stronger guarantees w.r.t. availability and performance characteristics.
For more information – please see: https://devblogs.microsoft.com/cosmosdb/serverless-preview/
Group By should support cross-partition queries.13 votes
This is supported currently. In which SDK version do you not see this?
I'd love to see more fine-grained permissions. In particular, an 'Edit' permission would be extremely helpful. The user with an Edit permission would be able to change any of the data for the resource except the ID, and would not be able to delete the document.
Use case: Provide users with direct but limited access to the DocumentDB database, rather than having to route all their requests through my own API.11 votes
At the moment, STRINGEQUALS with the case-insensitive flag requires a linear index scan over the property, so it is cardinality dependent, resulting in extremely high RU charges in large collections. Please implement a (perhaps optional) index that supports cardinality independent case-insensitive STRINGEQUALS. At the moment, the ugly workaround is to manually store duplicates of the property values in lower case. See also https://stackoverflow.com/questions/66024029/high-ru-charge-with-case-insensitive-search/66249495#662494959 votes
In some cases, string system functions may need to fall back to a scan. Adding an ORDER BY clause can help prevent this.
Please give this a try and let us know if this resolves the issue.
Currently I am using the GraphSONMode.Normal to get adjcency information of a 'vertex' but then I only use the outgoing edge information and ignore the incomming edge information.
From optimization point of view we should add more options to have more fine-grained control on the output format of the gremlin query.
I would like to suggect that we add 'IncludeOutE' option for GraphSONMode to return GraphSON format including only Outgoing edge information and 'IncludeInE' option to return GraphSON format including only Incomming edge information.7 votes
This is currently not on our road map. We’d like to get more feedback on this to see if there are more users that would want this. Will mark as need-feedback to see if this gets more votes and input from users.
Thank you for your suggestion.
Like other charts, Can we have an option to pin to dashboard for monitoring Throttled requests for cosmosdb7 votes
We currently have a new feature AppInsights in Preview. Please take a look and let us know if this meets your needs!
See attached for an example. The push pins allow you pin it there.
You should add more query tabs on CosmosDB for Gremlin, so that users can run different queries and compare the result without deleting every time the content of the textbox4 votes
Thank you for your suggestion. This is currently not on our roadmap but want to get feedback and input.
How many tabs do you think you need?
Please consider to add availability to deploy Cosmos DB with custom TYPE for Cassandra API TABLEs using ARM template.3 votes
Thank you for your suggestion.
Are you referring to custom data types here?
Or are you referring to user-defined types here?
If the emulator is started with /NoExplorer, then it will never open the data explorer even if explicitly requested to (i.e., right-click on icon, select "Open Data Explorer"). It would be better to have the /NoExplorer flag prevent the explorer from opening on startup, but allow it to be opened if explicitly requested.1 vote
Thanks Stephen for your suggestion.
The use case for this flag is for when emulator is used as part of CI/CD pipeline.
Is there a reason why this suggestion would be needed versus just stopping and starting the emulator again?
- Don't see your idea?