UAA api requests timeout when more than 200 users are requested per page
search cancel

UAA api requests timeout when more than 200 users are requested per page

book

Article ID: 422515

calendar_today

Updated On:

Products

VMware Tanzu Application Service

Issue/Introduction

While using cf-mgmt to sync users to cf and experience an issue since we updated to 10.3.1.

The problem is reproducible via cf cli 

When a get of 100 Users via the API works:

cf curl -v "/v3/users?per_page=100"            

RESPONSE: [2025-12-01T10:42:57+01:00]
HTTP/1.1 200 OK

However if the number of users per page is increased to 200 the following error is observed:

cf curl -v "/v3/users?per_page=200"     

REQUEST: [2025-12-01T10:52:10+01:00]
GET /v3/users?per_page=200 HTTP/1.1

RESPONSE: [2025-12-01T10:53:13+01:00]
HTTP/1.1 503 Service Unavailable
...
{
  "errors": [
    {
      "code": 20004,
      "detail": "The UAA service is currently unavailable",
      "title": "CF-UaaUnavailable"
    }
  ]
}

                                                                                                                                                                                

Test completed with about 600 users added in the Elastic Application Runtime

Environment

Elastic Application Runtime 10.3.1

Cause

An issue have been identified that the setting max-http-header-size was renamed to max-http-request-header-size in Spring Boot 3. Elastic Application Runtime 10.3.1 config was still setting the old value, causing unexpected response when a larger query is executed.

The fix  is expected in the January release in  Elastic Application Runtime 10.3.3

Resolution

Recommended action is to wait for version 10.3.3 to implement the fix, however if this issue is causing problems with the requests when manipulating users below steps can be completed to mitigate the problem:

This process have to be completed on each UAA instance.

 

bosh -d cf-<ID> ssh uaa/<INDEX>
#once logged in
sudo -i
#make a backup of file /var/vcap/data/uaa/boot/application.yml
cp /var/vcap/data/uaa/boot/application.yml /tmp/.
#edit the file /var/vcap/data/uaa/boot/application.yml and replace max-http-header-size with max-http-request-header-size
sudo sed -i 's/max-http-header-size/max-http-request-header-size/g' /var/vcap/data/uaa/boot/application.yml
#verify the change is complete
grep "max-http-request-header-size" /var/vcap/data/uaa/boot/application.yml

    max-http-request-header-size: 14336
  max-http-request-header-size: 14336
# reastart UAA service 
monit restart uaa
#verify Uaa service is running:
watch monit summary 

 

To confirm the operations completes:

cf curl -v "/v3/users?page=2&per_page=3000"

RESPONSE: [2025-12-16T09:46:47Z]
HTTP/1.1 200 OK

{
...
    "total_pages": 1,
    "total_results": 598
  },
  "resources": []
}

Additional Information

https://github.com/spring-projects/spring-boot/wiki/Spring-Boot-3.0.0-RC1-Release-Notes