CosmosDB container autoscale max throughput limited doesn't match documentation

Question

CosmosDB container autoscale max throughput limited doesn't match documentation

Josiah Vinson 20 Microsoft Employee

I'm attempting to update the autoscale maximum throughput on a Cosmos DB SQL container to 200,000 RU/s, which is well within the documented maximum of 1,000,000 RU/s.

However, the request fails with the following error:

Code: BadRequest Message: "The offer should have valid throughput values between 1000 and 100000 inclusive in increments of 1000. Requested throughput 200000 is more than possible maximum throughput 100000." ActivityId: 9d63f8a7-4592-11f1-b734-000d3afb4800

The discrepancy:

The error states the maximum possible throughput is 100,000 RU/s, but the official documentation at

https://aka.ms/max-provisioned-throughput indicates autoscale supports up to 1,000,000 RU/s (or higher by request).

Either:

The account/container has an undocumented quota that isn't surfaced in the portal, or
A quota increase is required to exceed 100K despite the docs suggesting it should be available by default

Request: Please clarify why this account is limited to 100,000 RU/s and either raise the limit to allow 200,000 RU/s or update

the documentation to reflect the actual default cap.

Josiah Vinson 20 Reputation points Microsoft Employee

2026-05-01T20:48:01.0866667+00:00
I was using an account that had no autoscale limit set, and I was able to:

Set the account autoscale throughput limit to 1M

Then set a container autoscale throughput limit to 100k.

This seems unintuitive - unlimited account throughput places a limit on how high the container throughput limit can be?
Josiah Vinson 20 Reputation points Microsoft Employee

2026-05-01T20:58:24.38+00:00

@Manoj Kumar Boyini I'm using an account with no limit for autoscale at the account level, and autoscale limits set at the container level. Does this limit the number of partitions available to each container?
Manoj Kumar Boyini 13,930 Reputation points Microsoft External Staff Moderator

2026-05-01T22:10:23.4633333+00:00

Hi @Josiah Vinson

The account-level autoscale setting does not limit the number of partitions per container, and partitions are not the reason for the 100K RU/s cap.

The limit you’re seeing is due to the current throughput capacity allocated to the account, which can be lower than the documented maximum of 1M RU/s.

To go beyond 100K RU/s, a throughput limit increase via support request is required. Once approved, you’ll be able to scale to 200K RU/s.
Manoj Kumar Boyini 13,930 Reputation points Microsoft External Staff Moderator

2026-05-04T06:04:36.0233333+00:00

Hi @Josiah Vinson

I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

Answer accepted by question author

0 additional answers

Your answer

Josiah Vinson 20 Reputation points Microsoft Employee

2026-05-01T20:48:01.0866667+00:00

I was using an account that had no autoscale limit set, and I was able to:

Set the account autoscale throughput limit to 1M

Then set a container autoscale throughput limit to 100k.

This seems unintuitive - unlimited account throughput places a limit on how high the container throughput limit can be?
Josiah Vinson 20 Reputation points Microsoft Employee

2026-05-01T20:58:24.38+00:00

@Manoj Kumar Boyini I'm using an account with no limit for autoscale at the account level, and autoscale limits set at the container level. Does this limit the number of partitions available to each container?
Manoj Kumar Boyini 13,930 Reputation points Microsoft External Staff Moderator

2026-05-01T22:10:23.4633333+00:00

Hi @Josiah Vinson

The account-level autoscale setting does not limit the number of partitions per container, and partitions are not the reason for the 100K RU/s cap.

The limit you’re seeing is due to the current throughput capacity allocated to the account, which can be lower than the documented maximum of 1M RU/s.

To go beyond 100K RU/s, a throughput limit increase via support request is required. Once approved, you’ll be able to scale to 200K RU/s.
Manoj Kumar Boyini 13,930 Reputation points Microsoft External Staff Moderator

2026-05-04T06:04:36.0233333+00:00

Hi @Josiah Vinson

I hope you had a chance to review the information shared earlier, and I hope this information has been helpful! If you still have questions, please let us know what is needed in the comments so the question can be answered.

Answer 1

Hi @Josiah Vinson

Based on your scenario, this behavior is due to the current maximum throughput supported by the container’s existing partition layout, rather than a discrepancy in the documented service limits.

While Azure Cosmos DB supports autoscale throughput up to 1,000,000 RU/s per container or database, this represents the service maximum, not the instantaneous limit for every resource. The value you can set at any given time depends on the container’s current capacity, which is determined by its physical partitions.

Each physical partition supports up to 10,000 RU/s, and the error indicating a maximum of 100,000 RU/s suggests that the container currently has capacity aligned with approximately 10 partitions. Until the service expands this partition layout, requests above this value (such as 200,000 RU/s) will be rejected.

This is expected behavior and not typically related to a quota restriction.

Next steps:

To proceed, you can:

Check the current maximum throughput (for example, via portal indicators such as instant maximum throughput)
Increase throughput in a way that allows the service to expand partitions (this may take some time depending on data and workload)
If you still require higher limits beyond the supported range, a support request can be raised for quota increase

For reference:
https://learn.microsoft.com/azure/cosmos-db/concepts-limits
https://learn.microsoft.com/azure/cosmos-db/scaling-provisioned-throughput-best-practices
https://aka.ms/max-provisioned-throughput

Share via

CosmosDB container autoscale max throughput limited doesn't match documentation

0 additional answers

Your answer