[AI Infra] Fix Observability AI assistant product docs missing multilingual support by qn895 · Pull Request #224274 · elastic/kibana

qn895 · 2025-06-17T14:42:04Z

Summary

This PR fixes #222176, and rewired the productDoc installation process to accept an inferenceId argument to the productDocBase installation API. It:

Allows for concurrent installation of the product docs with two different models: The default ELSER and the multilingual E5. Kibana will only install the one the user needs, but has capacity for other models if user needs both of them (i.e. ELSER for Security AI Assistant and multilingual E5 for Observability AI Assistant).
Modifies the script that generates the artifacts to also allows inferenceId to be passed in.

node scripts/build_product_doc_artifacts.js --product-name=security --stack-version=8.18  --inference-id=.multilingual-e5-small-elasticsearch

In parallel with this PR, deploys the public multilingual product doc artifacts for 8.18
It modifies the installation logic to append the inferenceId's to the target index name's (to distinguish it from the ELSER default) and define the mapping of the target index to use the E5's model_settings
Surfaces up error if there's an error with the installation

For example, if there's no corresponding artifact available, or if the artifact fails to fetch. Before:

After, it will prompt the user:

Note for Reviewers:

kibana-core: Saved object 'product-doc-install-status' was updated to add a new field inference_id
Security Gen AI: With the newly required inferenceId parameter to the installation endpoints, by default it will use ELSER '.elser-2-elasticsearch'
Observability AI Assistant: There are 2 at least todos remaining:
1. Make to sure pass the inferenceId to the retrieveDocumentation so that it reroutes to the right index https://github.com/elastic/kibana/pull/224274/files#diff-e393e350cf2449f8b756cad947fc8a902fddf6e6b30f1363750d469fc7d81b61R74
2. Handle the change in inference model selection for Product Doc. Here this is triggering an update to the product doc installation when user clicks Update model: https://github.com/elastic/kibana/pull/224274/files#diff-7d84fc1bf3106fe3b0cb357c800faefc1b96b853beeb74711f1c3c623ae901b9R151
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

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
...

…d-arg

…back to default ELSER. Surface up error if artifact not available with the inference.

qn895 · 2025-06-22T20:19:03Z

/ci

…-fix'

sorenlouv · 2025-06-23T11:01:16Z

...lugins/private/observability_ai_assistant_management/public/hooks/use_install_product_doc.ts

    [REACT_QUERY_KEYS.INSTALL_PRODUCT_DOC],
-    () => {
-      return productDocBase!.installation.install();
+    (inferenceId?: string) => {


Imo this should be required. Any scenarios where we don't want to pass inferenceId?

Suggested change

(inferenceId?: string) => {

(inferenceId: string) => {

Here it's because we will always fallback to ELSER by default, but we can make it a strict enforcement to pass in the inferenceId.

pgayvallet · 2025-06-23T11:08:53Z

x-pack/platform/plugins/shared/ai_infra/product_doc_base/server/tasks/ensure_up_to_date.ts

      id: ENSURE_DOC_UP_TO_DATE_TASK_ID,
      taskType: ENSURE_DOC_UP_TO_DATE_TASK_TYPE,
-      params: {},
+      params: { inferenceId },


AFAIK ensureScheduled tasks are uniquely scheduled by ID. Meaning that if you have different sources installing for different inferenceId at the same time, only one will be scheduled. Given security solution and the o11y assistant may call it at the same time at startup, this may be an issue

pgayvallet

So, I really don't think that allowing API consumers to directly control which inferenceId to use when installing the product doc is a good idea. This may lead to the product doc being installed multiple times for different models (e.g security solution and o11y assistant using different models on a deployment - which will occurs in the PR current state given given security solution will still use the default elser), and add a lot of complexity in the installation /management logic.

Why not instead having a single Kibana configuration property? That way, customers (or us during testing) can still control which model to use, but we're making sure there is only one single installation of the product doc (e.g security solution and o11y can't select different models). This seems to answer the initial need, while keeping everything more simple and resilient, WDYT?

sorenlouv · 2025-06-23T11:45:19Z

https://github.com/elastic/obs-ai-assistant-team/issues/300 is a follow-up issues that depends on the work in this PR.

sorenlouv · 2025-06-23T12:00:26Z

So, I really don't think that allowing API consumers to directly control which inferenceId to use when installing the product doc is a good idea. This may lead to the product doc being installed multiple times for different models (e.g security solution and o11y assistant using different models on a deployment - which will occurs in the PR current state given given security solution will still use the default elser), and add a lot of complexity in the installation /management logic.

Why not instead having a single Kibana configuration property? That way, customers (or us during testing) can still control which model to use, but we're making sure there is only one single installation of the product doc (e.g security solution and o11y can't select different models). This seems to answer the initial need, while keeping everything more simple and resilient, WDYT?

When the inferenceId for the knowledge base changes, we need to update the inferenceId for the product docs. How can we do that if the installProductDocs method doesn't accept inferenceId?

… not the default ELSER, installation status logic

sorenlouv · 2025-06-26T21:30:09Z

...lugins/shared/ai_infra/product_doc_base/public/services/installation/installation_service.ts

    if (!response.installed) {
-      throw new Error('Installation did not complete successfully');
+      throw new Error(
+        `Installation did not complete successfully.\n${response.failureReason ?? ''}`


I think customers will find this error quite confusing:

This is particularly cryptic:

End of central directory record signature not found. either not a zip file or file is truncated

What action are users expected to take from seeing this? And under what circumstances does it happen?

Yeah this is coming from cloud bucket where the artifact is missing. I can remove that message entirely.

sorenlouv · 2025-06-26T21:38:12Z

...ervability_ai_assistant_management/public/routes/components/settings_tab/change_kb_model.tsx


 export function ChangeKbModel({ knowledgeBase }: { knowledgeBase: UseKnowledgeBaseResult }) {
  const { overlays } = useKibana().services;
+  const { selectedInferenceId, setSelectedInferenceId } = useInferenceId();


Do we need useInferenceId? Can't we get the current inference id from knowledgeBase: knowledgeBase.status.value.currentInferenceId ?

sorenlouv · 2025-06-26T21:40:22Z

...vability_ai_assistant_management/public/routes/components/settings_tab/product_doc_entry.tsx


 export function ProductDocEntry() {
  const { overlays } = useKibana().services;
+  const { selectedInferenceId } = useInferenceId();


Can you use

Suggested change

const { selectedInferenceId } = useInferenceId();

const knowledgeBase = useKnowledgeBase();

and then knowledgeBase.status.value.currentInferenceId?

Initially this is what I did, but the UX was confusing. So when you switch to a different model, I wanted to show the product base for that inference is not installed yet. If we still use currently deployed the info is outdated. But happy to switch it back to your suggestion.

@sorenlouv Here's the difference for the install doc status. This is with the current approach with the useInferenceId

Screen.Recording.2025-06-27.at.12.59.50.mov

And this is with the knowledgeBase.status.value.currentInferenceId

Screen.Recording.2025-06-27.at.12.59.09.mov

I suggest going with const knowledgeBase = useKnowledgeBase();. It's the simpler option and I don't believe the UX is worse.

sorenlouv · 2025-06-27T09:11:56Z

❌ Manual testing failed

I tried running the command in the PR description:

node scripts/build_product_doc_artifacts.js --product-name=observability --stack-version=9.1.0  --inference-id=.multilingual-e5-small-elasticsearch

which returns

Missing required arguments: sourceClusterUrl, sourceClusterUsername, sourceClusterPassword, embeddingClusterUrl, embeddingClusterUsername, embeddingClusterPassword

Afterwards I tried installing the product docs via the UI. It just shows a spinner "Installing..." indefinitely. In the terminal I see these errors:

[2025-06-27T11:05:17.107+02:00][INFO ][plugins.productDocBase.doc-manager] Task ProductDocBase:InstallAllMultilingual scheduled to run soon
[2025-06-27T11:05:20.060+02:00][INFO ][plugins.productDocBase.package-installer] Starting installing documentation for product [kibana] and version [8.18] with inference ID [.multilingual-e5-small-elasticsearch]
[2025-06-27T11:05:20.632+02:00][ERROR][plugins.productDocBase.package-installer] Error during documentation installation of product [kibana]/[8.18] : No artifact available for product [kibana]/[8.18] for Inference ID [.multilingual-e5-small-elasticsearch]. End of central directory record signature not found. Either not a zip file, or file is truncated.
[2025-06-27T11:05:21.456+02:00][ERROR][plugins.taskManager] Task ProductDocBase:InstallAll "ProductDocBase:InstallAllMultilingual" failed: Error: End of central directory record signature not found. Either not a zip file, or file is truncated.

[2025-06-27T11:05:53.040+02:00][INFO ][plugins.productDocBase.package-installer] Starting installing documentation for product [kibana] and version [8.18] with inference ID [.multilingual-e5-small-elasticsearch]
[2025-06-27T11:05:54.842+02:00][ERROR][plugins.productDocBase.package-installer] Error during documentation installation of product [kibana]/[8.18] : No artifact available for product [kibana]/[8.18] for Inference ID [.multilingual-e5-small-elasticsearch]. End of central directory record signature not found. Either not a zip file, or file is truncated.
[2025-06-27T11:05:55.655+02:00][ERROR][plugins.taskManager] Task ProductDocBase:InstallAll "ProductDocBase:InstallAllMultilingual" failed: Error: End of central directory record signature not found. Either not a zip file, or file is truncated.

I assume the error is because I was unable to run the node script. Is that something I need to do because I am a dev and this is in development, or do we expect customers to run this command as well?

sorenlouv · 2025-06-27T11:13:01Z

I was able to test this after adding the following to kibana.yml:

xpack.productDocBase.artifactRepositoryUrl: https://storage.googleapis.com/kibana-ai-assistant-kb-artifacts-dev

darnautov · 2025-07-02T15:06:38Z

@elasticmachine merge upstream

stephmilovic

LGTM! Thanks for the fix

…apping'

…inference-id-arg

github-actions · 2025-07-03T07:50:47Z

🤖 GitHub comments

Expand to view the GitHub comments

Just comment with:

/oblt-deploy : Deploy a Kibana instance using the Observability test environments.
run docs-build : Re-trigger the docs validation. (use unformatted text in the comment!)

kibanamachine · 2025-07-03T09:53:30Z

Starting backport for target branches: 8.19, 9.1

https://github.com/elastic/kibana/actions/runs/16047278927

kibanamachine · 2025-07-03T09:53:45Z

Starting backport for target branches: 8.19, 9.1

https://github.com/elastic/kibana/actions/runs/16047284396

elasticmachine · 2025-07-03T09:54:25Z

💚 Build Succeeded

Buildkite Build
Commit: 1eecfe1

Metrics [docs]

Module Count

Fewer modules leads to a faster build time

id	before	after	diff
`automaticImport`	763	796	+33
`elasticAssistant`	418	451	+33
`productDocBase`	7	40	+33
`securitySolution`	7787	7820	+33
total			+132

Public APIs missing comments

Total count of every public API that lacks a comment. Target amount is 0. Run node scripts/build_api_docs --plugin [yourplugin] --stats comments for more detailed information.

id	before	after	diff
`@kbn/product-doc-common`	32	34	+2
`observabilityAIAssistant`	434	438	+4
total			+6

Async chunks

Total size of all lazy-loaded chunks that will be downloaded as the user navigates the app

id	before	after	diff
`observabilityAiAssistantManagement`	97.4KB	97.6KB	+193.0B
`securitySolution`	9.8MB	9.8MB	+69.0B
total			+262.0B

Page load bundle

Size of the bundles that are downloaded on every page load. Target size is below 100kb

id	before	after	diff
`productDocBase`	2.3KB	2.7KB	+387.0B

Saved Objects .kibana field count

Every field in each saved object type adds overhead to Elasticsearch. Kibana needs to keep the total field count below Elasticsearch's default limit of 1000 fields. Only specify field mappings for the fields you wish to search on or query. See https://www.elastic.co/guide/en/kibana/master/saved-objects-service.html#_mappings

id	before	after	diff
`product-doc-install-status`	6	7	+1

Unknown metric groups

API count

id	before	after	diff
`@kbn/product-doc-common`	32	34	+2
`llmTasks`	24	25	+1
`observabilityAIAssistant`	440	444	+4
`productDocBase`	23	24	+1
total			+8

History

💛 Build #315231 was flaky ff62d2d
💔 Build #315073 failed bddeeca
💔 Build #314949 failed c57ad07
💔 Build #314758 failed 7f29354
💚 Build #314194 succeeded 4102fd5

cc @qn895

kibanamachine · 2025-07-03T10:00:01Z

💔 All backports failed

Status	Branch	Result
❌	8.19	Backport failed because of merge conflicts
❌	9.1	Backport failed because of merge conflicts

Manual backport

To create the backport manually run:

node scripts/backport --pr 224274

Questions ?

Please refer to the Backport tool documentation

kibanamachine · 2025-07-03T10:00:22Z

💔 All backports failed

Status	Branch	Result
❌	8.19	Backport failed because of merge conflicts
❌	9.1	Backport failed because of merge conflicts

Manual backport

To create the backport manually run:

node scripts/backport --pr 224274

Questions ?

Please refer to the Backport tool documentation

…ingual support (elastic#224274) ## Summary This PR fixes elastic#222176, and rewired the productDoc installation process to accept an `inferenceId` argument to the productDocBase installation API. It: - Allows for concurrent installation of the product docs with two different models: The default ELSER and the multilingual E5. Kibana will only install the one the user needs, but has capacity for other models if user needs both of them (i.e. ELSER for Security AI Assistant and multilingual E5 for Observability AI Assistant). - Modifies the script that generates the artifacts to also allows inferenceId to be passed in. ``` node scripts/build_product_doc_artifacts.js --product-name=security --stack-version=8.18 --inference-id=.multilingual-e5-small-elasticsearch ``` - In parallel with this PR, deploys the public multilingual product doc artifacts for 8.18 - It modifies the installation logic to append the inferenceId's to the target index name's (to distinguish it from the ELSER default) and define the mapping of the target index to use the E5's model_settings - Surfaces up error if there's an error with the installation For example, if there's no corresponding artifact available, or if the artifact fails to fetch. Before: ![image](https://github.com/user-attachments/assets/ac9fe8db-a34e-4e67-8471-56e8f4520fdd) After, it will prompt the user: ![image](https://github.com/user-attachments/assets/ff18c6bd-0c20-4227-aac7-913a3032a31f) ## Note for Reviewers: - **kibana-core**: Saved object 'product-doc-install-status' was updated to add a new field `inference_id` - **Security Gen AI**: With the newly required inferenceId parameter to the installation endpoints, by default it will use ELSER '.elser-2-elasticsearch' - **Observability AI Assistant**: There are 2 at least todos remaining: 1) Make to sure pass the inferenceId to the retrieveDocumentation so that it reroutes to the right index https://github.com/elastic/kibana/pull/224274/files#diff-e393e350cf2449f8b756cad947fc8a902fddf6e6b30f1363750d469fc7d81b61R74 2) Handle the change in inference model selection for Product Doc. Here this is triggering an update to the product doc installation when user clicks Update model: https://github.com/elastic/kibana/pull/224274/files#diff-7d84fc1bf3106fe3b0cb357c800faefc1b96b853beeb74711f1c3c623ae901b9R151 - ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [x] 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) - [x] [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 - [ ] 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) ### 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) - [ ] ... --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Dima Arnautov <dmitrii.arnautov@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> (cherry picked from commit 4ff731d) # Conflicts: # x-pack/solutions/observability/plugins/observability_ai_assistant_app/server/functions/documentation.ts

darnautov · 2025-07-03T10:34:04Z

💚 All backports created successfully

Status	Branch	Result
✅	9.1
✅	8.19

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

Questions ?

Please refer to the Backport tool documentation

…ingual support (elastic#224274) ## Summary This PR fixes elastic#222176, and rewired the productDoc installation process to accept an `inferenceId` argument to the productDocBase installation API. It: - Allows for concurrent installation of the product docs with two different models: The default ELSER and the multilingual E5. Kibana will only install the one the user needs, but has capacity for other models if user needs both of them (i.e. ELSER for Security AI Assistant and multilingual E5 for Observability AI Assistant). - Modifies the script that generates the artifacts to also allows inferenceId to be passed in. ``` node scripts/build_product_doc_artifacts.js --product-name=security --stack-version=8.18 --inference-id=.multilingual-e5-small-elasticsearch ``` - In parallel with this PR, deploys the public multilingual product doc artifacts for 8.18 - It modifies the installation logic to append the inferenceId's to the target index name's (to distinguish it from the ELSER default) and define the mapping of the target index to use the E5's model_settings - Surfaces up error if there's an error with the installation For example, if there's no corresponding artifact available, or if the artifact fails to fetch. Before: ![image](https://github.com/user-attachments/assets/ac9fe8db-a34e-4e67-8471-56e8f4520fdd) After, it will prompt the user: ![image](https://github.com/user-attachments/assets/ff18c6bd-0c20-4227-aac7-913a3032a31f) ## Note for Reviewers: - **kibana-core**: Saved object 'product-doc-install-status' was updated to add a new field `inference_id` - **Security Gen AI**: With the newly required inferenceId parameter to the installation endpoints, by default it will use ELSER '.elser-2-elasticsearch' - **Observability AI Assistant**: There are 2 at least todos remaining: 1) Make to sure pass the inferenceId to the retrieveDocumentation so that it reroutes to the right index https://github.com/elastic/kibana/pull/224274/files#diff-e393e350cf2449f8b756cad947fc8a902fddf6e6b30f1363750d469fc7d81b61R74 2) Handle the change in inference model selection for Product Doc. Here this is triggering an update to the product doc installation when user clicks Update model: https://github.com/elastic/kibana/pull/224274/files#diff-7d84fc1bf3106fe3b0cb357c800faefc1b96b853beeb74711f1c3c623ae901b9R151 - ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [x] 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) - [x] [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 - [ ] 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) ### 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) - [ ] ... --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Dima Arnautov <dmitrii.arnautov@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com> (cherry picked from commit 4ff731d) # Conflicts: # src/core/server/integration_tests/ci_checks/saved_objects/check_registered_types.test.ts # x-pack/packages/ai-infra/product-doc-artifact-builder/src/build_artifacts.ts # x-pack/solutions/observability/plugins/observability_ai_assistant_app/scripts/evaluation/scenarios/documentation/index.spec.ts # x-pack/solutions/observability/plugins/observability_ai_assistant_app/server/functions/documentation.ts

…multilingual support (#224274) (#226368) # Backport This will backport the following commits from `main` to `9.1`: - [[AI Infra] Fix Observability AI assistant product docs missing multilingual support (#224274)](#224274)  ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sorenlouv/backport)  Co-authored-by: Quynh Nguyen (Quinn) <43350163+qn895@users.noreply.github.com>

… multilingual support (#224274) (#226378) # Backport This will backport the following commits from `main` to `8.19`: - [[AI Infra] Fix Observability AI assistant product docs missing multilingual support (#224274)](#224274)  ### Questions ? Please refer to the [Backport tool documentation](https://github.com/sorenlouv/backport)  --------- Co-authored-by: Quynh Nguyen (Quinn) <43350163+qn895@users.noreply.github.com>

…ingual support (elastic#224274) ## Summary This PR fixes elastic#222176, and rewired the productDoc installation process to accept an `inferenceId` argument to the productDocBase installation API. It: - Allows for concurrent installation of the product docs with two different models: The default ELSER and the multilingual E5. Kibana will only install the one the user needs, but has capacity for other models if user needs both of them (i.e. ELSER for Security AI Assistant and multilingual E5 for Observability AI Assistant). - Modifies the script that generates the artifacts to also allows inferenceId to be passed in. ``` node scripts/build_product_doc_artifacts.js --product-name=security --stack-version=8.18 --inference-id=.multilingual-e5-small-elasticsearch ``` - In parallel with this PR, deploys the public multilingual product doc artifacts for 8.18 - It modifies the installation logic to append the inferenceId's to the target index name's (to distinguish it from the ELSER default) and define the mapping of the target index to use the E5's model_settings - Surfaces up error if there's an error with the installation For example, if there's no corresponding artifact available, or if the artifact fails to fetch. Before: ![image](https://github.com/user-attachments/assets/ac9fe8db-a34e-4e67-8471-56e8f4520fdd) After, it will prompt the user: ![image](https://github.com/user-attachments/assets/ff18c6bd-0c20-4227-aac7-913a3032a31f) ## Note for Reviewers: - **kibana-core**: Saved object 'product-doc-install-status' was updated to add a new field `inference_id` - **Security Gen AI**: With the newly required inferenceId parameter to the installation endpoints, by default it will use ELSER '.elser-2-elasticsearch' - **Observability AI Assistant**: There are 2 at least todos remaining: 1) Make to sure pass the inferenceId to the retrieveDocumentation so that it reroutes to the right index https://github.com/elastic/kibana/pull/224274/files#diff-e393e350cf2449f8b756cad947fc8a902fddf6e6b30f1363750d469fc7d81b61R74 2) Handle the change in inference model selection for Product Doc. Here this is triggering an update to the product doc installation when user clicks Update model: https://github.com/elastic/kibana/pull/224274/files#diff-7d84fc1bf3106fe3b0cb357c800faefc1b96b853beeb74711f1c3c623ae901b9R151 - ### Checklist Check the PR satisfies following conditions. Reviewers should verify this PR satisfies this list as well. - [x] 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) - [x] [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 - [ ] 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) ### 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) - [ ] ... --------- Co-authored-by: kibanamachine <42973632+kibanamachine@users.noreply.github.com> Co-authored-by: Dima Arnautov <dmitrii.arnautov@elastic.co> Co-authored-by: Elastic Machine <elasticmachine@users.noreply.github.com>

qn895 added 3 commits June 17, 2025 08:20

Add new inference_id arg/param to install product doc endpoints

7d4cd90

Update to pass inferenceId through callback instead of hook

bf917d8

Update type

c7fa7fa

qn895 self-assigned this Jun 17, 2025

qn895 added the :ml label Jun 20, 2025

qn895 added 2 commits June 20, 2025 10:02

Merge remote-tracking branch 'upstream/main' into add-new-inference-i…

2327843

…d-arg

Add inference info

b9ae8d0

qn895 force-pushed the add-new-inference-id-arg branch from 7849fef to 079bdd5 Compare June 22, 2025 19:01

qn895 changed the title ~~[AI Infra] Add option to generate product docs with any Inference ID~~ [AI Infra] Add option to generate product docs with multilingual model E5 Jun 22, 2025

Update scripts to generate docs with optional inference_id, else fall…

132c0e6

…back to default ELSER. Surface up error if artifact not available with the inference.

qn895 force-pushed the add-new-inference-id-arg branch from 079bdd5 to 132c0e6 Compare June 22, 2025 20:18

[CI] Auto-commit changed files from 'node scripts/eslint --no-cache -…

380c589

…-fix'

sorenlouv reviewed Jun 23, 2025

View reviewed changes

pgayvallet reviewed Jun 23, 2025

View reviewed changes

qn895 added 5 commits June 25, 2025 11:52

Refactor to use another ID for E5

a0b0fd2

Update types

1508890

Fix default endpoint, types

f81b6ee

Fix get enpoint for status, update objects to reflect inference_id if…

026c8be

… not the default ELSER, installation status logic

Update tests

3fc9ab2

sorenlouv reviewed Jun 26, 2025

View reviewed changes

sorenlouv mentioned this pull request Jun 27, 2025

[Obs AI Assistant] Make product docs an implementation detail of the knowledge base #225580

Closed

qn895 added the bug Fixes for quality problems that affect the customer experience label Jun 27, 2025

darnautov requested a review from stephmilovic July 2, 2025 15:06

Merge branch 'main' into add-new-inference-id-arg

60c3370

stephmilovic approved these changes Jul 2, 2025

View reviewed changes

[CI] Auto-commit changed files from 'node scripts/styled_components_m…

bddeeca

…apping'

darnautov added the v9.1.0 label Jul 2, 2025

import defaultInferenceEndpoints from inference-common package

ff62d2d

darnautov added v8.19.0 and removed ci:project-deploy-observability Create an Observability project labels Jul 2, 2025

Merge remote-tracking branch 'upstream/main' into fork/qn895/add-new-…

1eecfe1

…inference-id-arg

botelastic bot added the ci:project-deploy-observability Create an Observability project label Jul 3, 2025

darnautov merged commit 4ff731d into elastic:main Jul 3, 2025
13 checks passed

kibanamachine added the v9.2.0 label Jul 3, 2025

darnautov mentioned this pull request Jul 3, 2025

[9.1] [AI Infra] Fix Observability AI assistant product docs missing multilingual support (#224274) #226368

Merged

kibanamachine mentioned this pull request Jul 3, 2025

[ES|QL] Improve binary expression grouping #225416

Merged

1 task

darnautov mentioned this pull request Jul 3, 2025

[8.19] [AI Infra] Fix Observability AI assistant product docs missing multilingual support (#224274) #226378

Merged

darnautov mentioned this pull request Jul 3, 2025

update adm-zip dependency to 0.5.16 #219079

Merged

6 tasks

sorenlouv mentioned this pull request Jul 4, 2025

[Obs AI Assistant] Upgrade testing for product docs with E5 #226524

Closed

	const { selectedInferenceId } = useInferenceId();
	const knowledgeBase = useKnowledgeBase();

Comments

Conversation

qn895 commented Jun 17, 2025 • edited by kibanamachine Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Note for Reviewers:

Checklist

Identify risks

Uh oh!

qn895 commented Jun 22, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

qn895 Jun 24, 2025 • 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!

Choose a reason for hiding this comment

Uh oh!

pgayvallet left a comment • 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!

sorenlouv commented Jun 23, 2025

Uh oh!

sorenlouv commented Jun 23, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

qn895 Jun 27, 2025 • 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!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

sorenlouv commented Jun 27, 2025

❌ Manual testing failed

Uh oh!

sorenlouv commented Jun 27, 2025

Uh oh!

darnautov commented Jul 2, 2025

Uh oh!

stephmilovic left a comment

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jul 3, 2025

🤖 GitHub comments

Uh oh!

Uh oh!

kibanamachine commented Jul 3, 2025

Uh oh!

kibanamachine commented Jul 3, 2025

Uh oh!

elasticmachine commented Jul 3, 2025

💚 Build Succeeded

Metrics [docs]

Module Count

Public APIs missing comments

Async chunks

Page load bundle

Saved Objects .kibana field count

API count

History

Uh oh!

kibanamachine commented Jul 3, 2025

💔 All backports failed

Manual backport

Questions ?

Uh oh!

kibanamachine commented Jul 3, 2025

💔 All backports failed

Manual backport

qn895 commented Jun 17, 2025 •

edited by kibanamachine

Loading

qn895 Jun 24, 2025 •

edited

Loading

pgayvallet left a comment •

edited

Loading

qn895 Jun 27, 2025 •

edited

Loading