[AI Connector][Inference endpoint] Creation: adaptive allocations not set correctly on serverless by saikatsarkar056 · Pull Request #251357 · elastic/kibana

saikatsarkar056 · 2026-02-02T22:12:20Z

Summary

Fixes AI Connector inference endpoint creation failing on serverless when using the 'elasticsearch' provider with adaptive allocations enabled.

The issue was that num_threads and num_allocations were being sent in the providerConfig alongside adaptive_allocations, which causes a conflict since these properties cannot be set when adaptive allocations is enabled.

This PR updates the form serializer to exclude num_threads and num_allocations from the payload when adaptive allocations is configured.

Screen.Recording.2026-02-02.at.2.58.32.PM.mov

Checklist

Check the PR satisfies following conditions.

Reviewers should verify this PR satisfies this list as well.

Any text added follows EUI's writing guidelines, uses sentence case text and includes i18n support
Documentation was added for features that require explanation or tutorials
Unit or functional tests were updated or added to match the most common scenarios
If a plugin configuration key changed, check if it needs to be allowlisted in the cloud and added to the docker list
This was checked for breaking HTTP API changes, and any breaking changes have been approved by the breaking-change committee. The release_note:breaking label should be applied in these situations.
Flaky Test Runner was used on any tests changed
The PR description includes the appropriate Release Notes section, and the correct release_note:* label is applied per the guidelines
Review the backport guidelines and apply applicable backport:* labels.

Identify risks

Does this PR introduce any risks? For example, consider risks like hard to test bugs, performance regression, potential of data loss.

Describe the risk, its severity, and mitigation for each identified risk. Invite stakeholders and evaluate how to proceed before merging.

See some risk examples
...

Release note

Fixes AI Connector creation on serverless failing when adaptive allocations is enabled with the ElasticSearch provider.

…location

saikatsarkar056 · 2026-02-10T16:50:06Z

This PR is blocked due to an backend error. Inference team is looking into the issue now. Once fixed, we can test the UI changes.

saikatsarkar056 · 2026-02-19T16:07:51Z

@elasticmachine merge upstream

…location

alvarezmelissa87 · 2026-02-20T18:43:05Z

...packages/shared/kbn-inference-endpoint-ui-common/src/components/inference_flyout_wrapper.tsx

              }
-            : {}),
+            : {
+                ...(numAllocations !== undefined ? { num_allocations: numAllocations } : {}),


Just for clarity - why is this change to the serializer needed?
This is explicitly checking for undefined - is it possible for numAllocations to be null?

When editing an endpoint on serverless, num_allocations was inside providerConfig and was being sent in the update request. However, Elasticsearch does not allow this field on serverless, so we need to remove it.

The cleanest solution is to remove (destructure) num_allocations from providerConfig at the beginning. After doing that, restProviderConfig will no longer include it.

For serverless, this is correct — the field should not be sent.
For stateful deployments, we still need num_allocations, so we add it back explicitly.

is it possible for numAllocations to be null?

Yes. There is no num_allocations field for OpenAI

alvarezmelissa87 · 2026-02-20T18:46:33Z

x-pack/platform/plugins/shared/inference_endpoint/server/routes/index.ts

-              ...(config?.providerConfig?.num_allocations !== undefined && {
-                num_allocations: config.providerConfig.num_allocations,
-              }),
+              ...(adaptiveAllocations


For ease of understanding and maintenance, we should follow the kibana styleguide and not use nested ternary operators.

alvarezmelissa87 · 2026-02-20T18:48:27Z

x-pack/platform/plugins/shared/inference_endpoint/server/routes/index.ts

-              }),
+              ...(adaptiveAllocations
+                ? { adaptive_allocations: adaptiveAllocations }
+                : numAllocations !== undefined


For this numAllocations undefined check - is it possible that it could be null?

It can be null for OpenAI

…location

saikatsarkar056 · 2026-02-23T16:57:08Z

@elasticmachine merge upstream

alvarezmelissa87

Code LGTM ⚡

saikatsarkar056 · 2026-02-23T18:26:21Z

@elasticmachine merge upstream

elasticmachine · 2026-02-23T18:40:24Z

💚 Build Succeeded

Buildkite Build
Commit: 6f931a1

History

💔 Build #399454 failed af24f72
💔 Build #398778 failed 8e4dbff
💛 Build #397949 was flaky d042e43
💔 Build #392192 failed 7eb33a6
💛 Build #390760 was flaky e6a5b66

cc @saikatsarkar056

kibanamachine · 2026-02-23T19:05:38Z

Starting backport for target branches: 9.3

https://github.com/elastic/kibana/actions/runs/22320695468

kibanamachine · 2026-02-23T19:13:23Z

💔 All backports failed

Status	Branch	Result
❌	9.3	Backport failed because of merge conflicts

Manual backport

To create the backport manually run:

node scripts/backport --pr 251357

Questions ?

Please refer to the Backport tool documentation

saikatsarkar056 · 2026-02-23T19:32:59Z

💚 All backports created successfully

Status	Branch	Result
✅	9.3

Note: Successful backport PRs will be merged automatically after passing CI.

Questions ?

Please refer to the Backport tool documentation

… set correctly on serverless (elastic#251357) ## Summary Fixes AI Connector inference endpoint creation failing on serverless when using the 'elasticsearch' provider with adaptive allocations enabled. The issue was that `num_threads` and `num_allocations` were being sent in the `providerConfig` alongside `adaptive_allocations`, which causes a conflict since these properties cannot be set when adaptive allocations is enabled. This PR updates the form serializer to exclude `num_threads` and `num_allocations` from the payload when adaptive allocations is configured. https://github.com/user-attachments/assets/5d4b42ff-db33-4f3a-ab70-b32fcccbc445 ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [ ] Any text added follows [EUI's writing guidelines](https://elastic.github.io/eui/#/guidelines/writing), uses sentence case text and includes [i18n support](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md) - [ ] [Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html) was added for features that require explanation or tutorials - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios - [ ] If a plugin configuration key changed, check if it needs to be allowlisted in the cloud and added to the [docker list](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker) - [ ] This was checked for breaking HTTP API changes, and any breaking changes have been approved by the breaking-change committee. The `release_note:breaking` label should be applied in these situations. - [ ] [Flaky Test Runner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1) was used on any tests changed - [x] The PR description includes the appropriate Release Notes section, and the correct `release_note:*` label is applied per the [guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process) - [x] Review the [backport guidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing) and apply applicable `backport:*` labels. ### Identify risks Does this PR introduce any risks? For example, consider risks like hard to test bugs, performance regression, potential of data loss. Describe the risk, its severity, and mitigation for each identified risk. Invite stakeholders and evaluate how to proceed before merging. - [ ] [See some risk examples](https://github.com/elastic/kibana/blob/main/RISK_MATRIX.mdx) - [ ] ... ### Release note Fixes AI Connector creation on serverless failing when adaptive allocations is enabled with the ElasticSearch provider. --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Melissa <melissa.alvarez@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> (cherry picked from commit 82dc85e) # Conflicts: # x-pack/platform/packages/shared/kbn-inference-endpoint-ui-common/src/components/inference_flyout_wrapper.test.tsx

… set correctly on serverless (elastic#251357) ## Summary Fixes AI Connector inference endpoint creation failing on serverless when using the 'elasticsearch' provider with adaptive allocations enabled. The issue was that `num_threads` and `num_allocations` were being sent in the `providerConfig` alongside `adaptive_allocations`, which causes a conflict since these properties cannot be set when adaptive allocations is enabled. This PR updates the form serializer to exclude `num_threads` and `num_allocations` from the payload when adaptive allocations is configured. https://github.com/user-attachments/assets/5d4b42ff-db33-4f3a-ab70-b32fcccbc445 ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [ ] Any text added follows [EUI's writing guidelines](https://elastic.github.io/eui/#/guidelines/writing), uses sentence case text and includes [i18n support](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md) - [ ] [Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html) was added for features that require explanation or tutorials - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios - [ ] If a plugin configuration key changed, check if it needs to be allowlisted in the cloud and added to the [docker list](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker) - [ ] This was checked for breaking HTTP API changes, and any breaking changes have been approved by the breaking-change committee. The `release_note:breaking` label should be applied in these situations. - [ ] [Flaky Test Runner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1) was used on any tests changed - [x] The PR description includes the appropriate Release Notes section, and the correct `release_note:*` label is applied per the [guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process) - [x] Review the [backport guidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing) and apply applicable `backport:*` labels. ### Identify risks Does this PR introduce any risks? For example, consider risks like hard to test bugs, performance regression, potential of data loss. Describe the risk, its severity, and mitigation for each identified risk. Invite stakeholders and evaluate how to proceed before merging. - [ ] [See some risk examples](https://github.com/elastic/kibana/blob/main/RISK_MATRIX.mdx) - [ ] ... ### Release note Fixes AI Connector creation on serverless failing when adaptive allocations is enabled with the ElasticSearch provider. --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Melissa <melissa.alvarez@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> (cherry picked from commit 82dc85e) Co-authored-by: Cursor <cursoragent@cursor.com> # Conflicts: # x-pack/platform/packages/shared/kbn-inference-endpoint-ui-common/src/components/inference_flyout_wrapper.test.tsx

…ns not set correctly on serverless (#251357) (#254585) # Backport This will backport the following commits from `main` to `9.3`: - [[AI Connector][Inference endpoint] Creation: adaptive allocations not set correctly on serverless (#251357)](#251357)  ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sorenlouv/backport)

… set correctly on serverless (elastic#251357) ## Summary Fixes AI Connector inference endpoint creation failing on serverless when using the 'elasticsearch' provider with adaptive allocations enabled. The issue was that `num_threads` and `num_allocations` were being sent in the `providerConfig` alongside `adaptive_allocations`, which causes a conflict since these properties cannot be set when adaptive allocations is enabled. This PR updates the form serializer to exclude `num_threads` and `num_allocations` from the payload when adaptive allocations is configured. https://github.com/user-attachments/assets/5d4b42ff-db33-4f3a-ab70-b32fcccbc445 ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [ ] Any text added follows [EUI's writing guidelines](https://elastic.github.io/eui/#/guidelines/writing), uses sentence case text and includes [i18n support](https://github.com/elastic/kibana/blob/main/src/platform/packages/shared/kbn-i18n/README.md) - [ ] [Documentation](https://www.elastic.co/guide/en/kibana/master/development-documentation.html) was added for features that require explanation or tutorials - [x] [Unit or functional tests](https://www.elastic.co/guide/en/kibana/master/development-tests.html) were updated or added to match the most common scenarios - [ ] If a plugin configuration key changed, check if it needs to be allowlisted in the cloud and added to the [docker list](https://github.com/elastic/kibana/blob/main/src/dev/build/tasks/os_packages/docker_generator/resources/base/bin/kibana-docker) - [ ] This was checked for breaking HTTP API changes, and any breaking changes have been approved by the breaking-change committee. The `release_note:breaking` label should be applied in these situations. - [ ] [Flaky Test Runner](https://ci-stats.kibana.dev/trigger_flaky_test_runner/1) was used on any tests changed - [x] The PR description includes the appropriate Release Notes section, and the correct `release_note:*` label is applied per the [guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process) - [x] Review the [backport guidelines](https://docs.google.com/document/d/1VyN5k91e5OVumlc0Gb9RPa3h1ewuPE705nRtioPiTvY/edit?usp=sharing) and apply applicable `backport:*` labels. ### Identify risks Does this PR introduce any risks? For example, consider risks like hard to test bugs, performance regression, potential of data loss. Describe the risk, its severity, and mitigation for each identified risk. Invite stakeholders and evaluate how to proceed before merging. - [ ] [See some risk examples](https://github.com/elastic/kibana/blob/main/RISK_MATRIX.mdx) - [ ] ... ### Release note Fixes AI Connector creation on serverless failing when adaptive allocations is enabled with the ElasticSearch provider. --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Melissa <melissa.alvarez@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>

saikatsarkar056 added 2 commits February 2, 2026 15:04

Fix adaptive allocations

669a356

Refactor the tests

07353f6

saikatsarkar056 requested review from a team as code owners February 2, 2026 22:12

saikatsarkar056 mentioned this pull request Feb 2, 2026

[AI Connector][Inference endpoint] Update: adaptive allocations not set correctly on serverless #251310

Closed

saikatsarkar056 changed the title ~~Fix adaptive allocation~~ [AI Connector][Inference endpoint] Creation: adaptive allocations not set correctly on serverless Feb 2, 2026

saikatsarkar056 self-assigned this Feb 2, 2026

saikatsarkar056 added backport:version Backport to applied version labels release_note:fix Team:Search v9.3.0 v9.4.0 labels Feb 2, 2026

kibanamachine and others added 3 commits February 2, 2026 22:28

Changes from yarn openapi:bundle

4328fdf

only remove num_allocations if max_allocations is set

4de0376

Fix the allocation edit issue

4b07baf

saikatsarkar056 requested a review from a team as a code owner February 3, 2026 00:39

saikatsarkar056 added 2 commits February 3, 2026 17:20

Restore the changes in bundle.json

76fcc48

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

2903ff7

…location

alvarezmelissa87 mentioned this pull request Feb 5, 2026

[ML] AI/Inference Connector creation: use 'location' field to correctly set provider config #250838

Merged

8 tasks

saikatsarkar056 added 5 commits February 5, 2026 17:05

Fix the test

e6a5b66

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

b21d9e3

…location

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

f4a377e

…location

Apply changes to only deserializer

3cfd70a

Refactor code

7eb33a6

peteharverson requested a review from alvarezmelissa87 February 10, 2026 16:30

saikatsarkar056 added the blocked label Feb 10, 2026

alvarezmelissa87 requested a review from Samiul-TheSoccerFan February 19, 2026 16:05

saikatsarkar056 added 4 commits February 20, 2026 11:03

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

f62ac6f

…location

Refactor code

5016080

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

57a399a

…location

Use omit for cleaner code

fab1afe

alvarezmelissa87 reviewed Feb 20, 2026

View reviewed changes

alvarezmelissa87 requested changes Feb 20, 2026

View reviewed changes

alvarezmelissa87 reviewed Feb 20, 2026

View reviewed changes

saikatsarkar056 added 3 commits February 20, 2026 16:21

Refactor code

8e4dbff

Refactor code

a2d537a

Merge branch 'main' of github.com:elastic/kibana into fix_adaptive_al…

9648bf2

…location

Merge branch 'main' into fix_adaptive_allocation

af24f72

saikatsarkar056 enabled auto-merge (squash) February 23, 2026 17:00

alvarezmelissa87 approved these changes Feb 23, 2026

View reviewed changes

Merge branch 'main' into fix_adaptive_allocation

6f931a1

saikatsarkar056 merged commit 82dc85e into elastic:main Feb 23, 2026
16 checks passed

saikatsarkar056 removed the blocked label Feb 23, 2026

saikatsarkar056 mentioned this pull request Feb 23, 2026

[9.3] [AI Connector][Inference endpoint] Creation: adaptive allocations not set correctly on serverless (#251357) #254585

Merged

kibanamachine added the v9.3.1 label Feb 23, 2026

Conversation

saikatsarkar056 commented Feb 2, 2026 • edited by kibanamachine Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Checklist

Identify risks

Release note

Uh oh!

saikatsarkar056 commented Feb 10, 2026

Uh oh!

saikatsarkar056 commented Feb 19, 2026

Uh oh!

alvarezmelissa87 Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 Feb 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 Feb 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alvarezmelissa87 Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

alvarezmelissa87 Feb 20, 2026

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 Feb 21, 2026

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 commented Feb 23, 2026

Uh oh!

alvarezmelissa87 left a comment

Choose a reason for hiding this comment

Uh oh!

saikatsarkar056 commented Feb 23, 2026

Uh oh!

elasticmachine commented Feb 23, 2026

💚 Build Succeeded

History

Uh oh!

Uh oh!

kibanamachine commented Feb 23, 2026

Uh oh!

kibanamachine commented Feb 23, 2026

💔 All backports failed

Manual backport

Questions ?

Uh oh!

saikatsarkar056 commented Feb 23, 2026

💚 All backports created successfully

Questions ?

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

saikatsarkar056 commented Feb 2, 2026 •

edited by kibanamachine

Loading

saikatsarkar056 Feb 20, 2026 •

edited

Loading

saikatsarkar056 Feb 20, 2026 •

edited

Loading