Kendra Create Index is stuck for over a day

Question

Hi,

I have been exploring Kendra lately, and my aws knowledge and hands on skills is intermediate to advance.

so last 2 days I just did the first step to create index and i chose enterprise_edition gen ai which was just the initial step to get my self goin. noticed it takes a while to create, so decided to create another with with developer edition, it was created less than 2 minutes, and i proceed to step 2 and 3, for web crawler as data source and monitor my report and logs in cloud watch they all were successful, apparently the very first that i started to create before hand with enterprise_edition was stuck, i began to doubt the 2nd day coz it was 27 hrs in "Creating" state, considering i havent put any data source yet. it's just the index being created which does not make sense. and I think this is an AWS Bug.

I attempted to force delete the index via aws cli in cloud shell, but it was helpless coz its in creating state and there is no other way to touch it, but wait for it to finish or atleast timeout which was not even an option.

Please see my logs:

~ $ aws kendra list-indices
{
    "IndexConfigurationSummaryItems": [
        {
            "Name": "intelligent-search-01",
            "Id": "1c24593a-3054-4297-9b18-516bf2f0ec16",
            "Edition": "GEN_AI_ENTERPRISE_EDITION",
            "CreatedAt": "2025-05-14T11:30:54.534000+00:00",
            "UpdatedAt": "2025-05-14T11:30:54.534000+00:00",
            "Status": "CREATING"
        },
        {
            "Name": "test-index",
            "Id": "3c2fae0d-196b-462c-abb2-425d4b858435",
            "Edition": "DEVELOPER_EDITION",
            "CreatedAt": "2025-05-15T09:36:02.761000+00:00",
            "UpdatedAt": "2025-05-15T09:43:48.376000+00:00",
            "Status": "ACTIVE"
        },
        {
            "Name": "sample-index",
            "Id": "7ba39f91-30c2-4b1d-8f46-67540e6dd1cf",
            "Edition": "DEVELOPER_EDITION",
            "CreatedAt": "2025-05-15T08:22:20.454000+00:00",
            "UpdatedAt": "2025-05-15T08:30:17.157000+00:00",
            "Status": "ACTIVE"
        }
    ]
}
~ $

~ $    aws logs describe-log-groups --log-group-name-prefix "/aws/kendra"

{
    "logGroups": [
        {
            "logGroupName": "/aws/kendra/3c2fae0d-196b-462c-abb2-425d4b858435",
            "creationTime": 1747302505451,
            "metricFilterCount": 0,
            "arn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/3c2fae0d-196b-462c-abb2-425d4b858435:*",
            "storedBytes": 0,
            "logGroupClass": "STANDARD",
            "logGroupArn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/3c2fae0d-196b-462c-abb2-425d4b858435"
        },
        {
            "logGroupName": "/aws/kendra/68c3b6d0-f8ad-429c-8709-16632eda90c4",
            "creationTime": 1747293860704,
            "metricFilterCount": 0,
            "arn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/68c3b6d0-f8ad-429c-8709-16632eda90c4:*",
            "storedBytes": 0,
            "logGroupClass": "STANDARD",
            "logGroupArn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/68c3b6d0-f8ad-429c-8709-16632eda90c4"
        },
        {
            "logGroupName": "/aws/kendra/7ba39f91-30c2-4b1d-8f46-67540e6dd1cf",
            "creationTime": 1747298085556,
            "metricFilterCount": 0,
            "arn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/7ba39f91-30c2-4b1d-8f46-67540e6dd1cf:*",
            "storedBytes": 0,
            "logGroupClass": "STANDARD",
            "logGroupArn": "arn:aws:logs:ap-southeast-2:066926217034:log-group:/aws/kendra/7ba39f91-30c2-4b1d-8f46-67540e6dd1cf"
        }
    ]
}
~ $ 
~ $    aws service-quotas list-service-quotas --service-code kendra
{
    "Quotas": []
}
~ $

~ $    aws kendra delete-index --id 1c24593a-3054-4297-9b18-516bf2f0ec16

An error occurred (ConflictException) when calling the DeleteIndex operation: Customer 066926217034 LogicalIndex 1c24593a-3054-4297-9b18-516bf2f0ec16 is in state CREATING at the moment, thus cannot be deleted
~ $

I just hoped it would have timed out instead, coz i can see AWS have not thought about situations like this, where end-user does not have control anymore to stop emergency stuck process even cli level, so that it will not have unnecessary bills charges or any cost implication for indexing hours, its past 36 hrs now almost 40 hours in creating state.

The problem with support is that, it requires them to upsell to me a business support subscription in order just to stop this, which is unbelievable, we rarely reach support and often have our manpower capacity with pools of expertise in technicalities. in this case Kendra I believe is not a frequently used services, therefore we had certain degree of credits to explore the service, however it got stuck on Enterprise Edition. although Developers Edition workflows were seamless. the problem here is that we are in a deadend state where no other solution from end-user perspective is viable.

Have anyone experience same situation?

Take note. There are no data source nor any job attached to this. its just the very preliminary stage where it got stuck.

re:Post Agent · Answer

I understand your frustration with the Kendra index creation being stuck for an extended period. This is indeed an unusual situation, especially for an index in the initial creation stage without any data sources attached.

For Amazon Kendra, the index creation process typically doesn't take more than a few hours, even for Enterprise Edition indexes. The fact that your index has been in the "CREATING" state for over 36 hours suggests there might be an underlying issue with the service or your specific index.

Unfortunately, as you've discovered, there's no direct way for users to forcibly stop or delete an index that's in the "CREATING" state. This is a limitation of the service designed to prevent data loss or corruption during the creation process.

In situations like this, where normal operations and CLI commands are not resolving the issue, the best course of action is typically to contact AWS Support. However, I understand your reluctance to upgrade to Business Support just to resolve this issue.

As an alternative, you could try the following:

1. Check if there are any Service Health Dashboard announcements or Personal Health Dashboard notifications related to Kendra in your region. This could indicate if there's a known service issue.

2. Monitor your AWS bill closely to ensure you're not being charged for the stuck index. Typically, you should only be billed for active indexes.

3. Continue to monitor the index status. While unlikely, it's possible the index could eventually time out or complete on its own.

4. If the issue persists, you might want to consider posting about this on the AWS Developer Forums or AWS re:Post. While these aren't official support channels, AWS employees sometimes monitor these forums and might be able to provide guidance or escalate the issue internally.

5. As a last resort, if you're concerned about potential charges or if this is blocking critical work, you might need to weigh the cost of a temporary upgrade to Business Support against the potential impact of leaving this unresolved.

This situation highlights a gap in the user control over Kendra resources, especially in error scenarios. It would be beneficial if AWS implemented a timeout or user-initiated cancellation option for stuck index creations in the future.  
**Sources**  
[Kendra GenAI index not available | AWS re:Post](https://repost.aws/questions/QUX-ZSKuVUSoqedagSO_RLxw/kendra-genai-index-not-available)  
[Creating an index - Amazon Kendra](https://docs.aws.amazon.com/kendra/latest/dg/create-index.html)  
[Amazon Kendra examples using AWS CLI - AWS SDK Code Examples](https://docs.aws.amazon.com/code-library/latest/ug/cli_2_kendra_code_examples.html)

mafthegreat · Answer

so indeed it was an AWS Issue.

I had to hit the button for 2 months trial for business support tier to speed things up.

It just kind of sucks that there is no contingency plan for emergency cases for the stuck resources.

as per support from business support tier subscription:

Our investigation confirmed that "This is a known issue, we've been having capacity issues in SYD, we'd advise using other regions to unblock customer, we are working on mitigating this in the meantime"

Kendra Create Index is stuck for over a day

Add your answer

Relevant content