Skip to content

[Obs AI Assistant] Check for documents before starting semantic text migration#221152

Merged
viduni94 merged 3 commits intoelastic:mainfrom
viduni94:check-for-documents-before-starting-semantic-text-migration
May 21, 2025
Merged

[Obs AI Assistant] Check for documents before starting semantic text migration#221152
viduni94 merged 3 commits intoelastic:mainfrom
viduni94:check-for-documents-before-starting-semantic-text-migration

Conversation

@viduni94
Copy link
Contributor

@viduni94 viduni94 commented May 21, 2025 •
edited by kibanamachine
Loading

Closes #221157

Summary

We run a semantic text migration at startup to add the semantic text field to documents that were created before 8.17.

Before multilingual KB was introduced:

  • We created index assets for KB when the AI Assistant flyout opens.
  • Even if the user does not set up the KB, they will have a component template pointing to the custom inference endpoint.

With the introduction of multilingual KB:

  • We moved some of the index creation to when setting up the KB.
  • We try to do the semantic_text migration at startup. During this migration, for users who didn't set up the KB but had the index assets created at startup, the custom inference endpoint will be unavailable.
  • But since the migration uses the inference endpoint from the write index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.

Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]

There is no customer impact from this, just that the error that gets logged is creating a lot of noise.

Solution

This PR checks whether there are documents without semantic_text before starting the migration.
And also reduced the log level to warn because we hit the /status endpoint once when a user opens the AI Assistant.

Checklist

  • The PR description includes the appropriate Release Notes section, and the correct release_note:* label is applied per the guidelines

@viduni94 viduni94 requested a review from a team as a code owner May 21, 2025 16:29
@botelastic botelastic bot added ci:project-deploy-observability Create an Observability project Team:Obs AI Assistant Observability AI Assistant labels May 21, 2025
@elasticmachine
Copy link
Contributor

elasticmachine commented May 21, 2025

Pinging @elastic/obs-ai-assistant (Team:Obs AI Assistant)

@github-actions
Copy link
Contributor

github-actions bot commented May 21, 2025

🤖 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!)

@viduni94 viduni94 self-assigned this May 21, 2025
@elasticmachine
Copy link
Contributor

elasticmachine commented May 21, 2025 •
edited
Loading

💛 Build succeeded, but was flaky

  • Buildkite Build
  • Commit: c4f5fe4
  • Kibana Serverless Image: docker.elastic.co/kibana-ci/kibana-serverless:pr-221152-c4f5fe4e4169

Failed CI Steps

Test Failures

  • [job] [logs] FTR Configs #119 / Cloud Security Posture Vulnerabilities Page - Grouping Default Grouping groups vulnerabilities by cloud account and sort by number of vulnerabilities desc
  • [job] [logs] FTR Configs #32 / ESQL execution logic API @ess @serverless ES|QL rule type alerts enrichment should be enriched with host risk score

Metrics [docs]

✅ unchanged

History

cc @viduni94

@viduni94
Copy link
Contributor Author

viduni94 commented May 21, 2025

I did some upgrade testing as well with documents and with no documents.
Upgrade path for ES - 8.16 --> 8.19 --> latest (main)

image

@viduni94 viduni94 merged commit b1e7012 into elastic:main May 21, 2025
10 checks passed
@kibanamachine
Copy link
Contributor

kibanamachine commented May 21, 2025

Starting backport for target branches: 8.19

https://github.com/elastic/kibana/actions/runs/15172778823

kibanamachine pushed a commit to kibanamachine/kibana that referenced this pull request May 21, 2025
@viduni94
[Obs AI Assistant] Check for documents before starting semantic text …
1b3689d 
…migration (elastic#221152)

Closes elastic#221157

## Summary

We run a semantic text migration at startup to add the semantic text
field to documents that were created before 8.17.

Before multilingual KB was introduced:
- We created index assets for KB when the AI Assistant flyout opens.
- Even if the user does not set up the KB, they will have a component
template pointing to the custom inference endpoint.

With the introduction of multilingual KB:
- We moved some of the index creation to when setting up the KB.
- We try to do the semantic_text migration at startup. During this
migration, for users who didn't set up the KB but had the index assets
created at startup, the custom inference endpoint will be unavailable.
- But since the migration uses the inference endpoint from the write
index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.
```
Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]
```

There is no customer impact from this, just that the error that gets
logged is creating a lot of noise.

## Solution

This PR checks whether there are documents without semantic_text before
starting the migration.
And also reduced the log level to warn because we hit the `/status`
endpoint once when a user opens the AI Assistant.

### Checklist

- [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)

(cherry picked from commit b1e7012)
@kibanamachine kibanamachine mentioned this pull request May 21, 2025
Merged
@kibanamachine
Copy link
Contributor

kibanamachine commented May 21, 2025

💚 All backports created successfully

Status Branch Result
✅ 8.19

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

Questions ?

Please refer to the Backport tool documentation

kibanamachine added a commit that referenced this pull request May 21, 2025
@kibanamachine @viduni94
[8.19] [Obs AI Assistant] Check for documents before starting semanti…
5ef460d 
…c text migration (#221152) (#221203)

# Backport

This will backport the following commits from `main` to `8.19`:
- [[Obs AI Assistant] Check for documents before starting semantic text
migration (#221152)](#221152)

<!--- Backport version: 9.6.6 -->

### Questions ?
Please refer to the [Backport tool
documentation](https://github.com/sorenlouv/backport)

<!--BACKPORT [{"author":{"name":"Viduni
Wickramarachchi","email":"viduni.wickramarachchi@elastic.co"},"sourceCommit":{"committedDate":"2025-05-21T21:18:06Z","message":"[Obs
AI Assistant] Check for documents before starting semantic text
migration (#221152)\n\nCloses
https://github.com/elastic/kibana/issues/221157\n\n## Summary\n\nWe run
a semantic text migration at startup to add the semantic text\nfield to
documents that were created before 8.17.\n\nBefore multilingual KB was
introduced:\n- We created index assets for KB when the AI Assistant
flyout opens.\n- Even if the user does not set up the KB, they will have
a component\ntemplate pointing to the custom inference endpoint.\n\nWith
the introduction of multilingual KB:\n- We moved some of the index
creation to when setting up the KB. \n- We try to do the semantic_text
migration at startup. During this\nmigration, for users who didn't set
up the KB but had the index assets\ncreated at startup, the custom
inference endpoint will be unavailable.\n- But since the migration uses
the inference endpoint from the write\nindex, we try to access an
endpoint that's not available.\n\nThis is the reason for this error to
be logged.\n```\nInference endpoint \"obs_ai_assistant_kb_inference\"
not found or unavailable: resource_not_found_exception\n\tRoot
causes:\n\t\tresource_not_found_exception: Inference endpoint not found
[obs_ai_assistant_kb_inference]\n```\n\nThere is no customer impact from
this, just that the error that gets\nlogged is creating a lot of
noise.\n\n## Solution\n\nThis PR checks whether there are documents
without semantic_text before\nstarting the migration.\nAnd also reduced
the log level to warn because we hit the `/status`\nendpoint once when a
user opens the AI Assistant.\n\n\n### Checklist\n\n- [x] The PR
description includes the appropriate Release Notes section,\nand the
correct `release_note:*` label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"b1e7012477af991632a0ad2fdfccad743bcac3c6","branchLabelMapping":{"^v9.1.0$":"main","^v(\\d+).(\\d+).\\d+$":"$1.$2"}},"sourcePullRequest":{"labels":["release_note:fix","Team:Obs
AI
Assistant","ci:project-deploy-observability","backport:version","v9.1.0","v8.19.0"],"title":"[Obs
AI Assistant] Check for documents before starting semantic text
migration","number":221152,"url":"https://github.com/elastic/kibana/pull/221152","mergeCommit":{"message":"[Obs
AI Assistant] Check for documents before starting semantic text
migration (#221152)\n\nCloses
https://github.com/elastic/kibana/issues/221157\n\n## Summary\n\nWe run
a semantic text migration at startup to add the semantic text\nfield to
documents that were created before 8.17.\n\nBefore multilingual KB was
introduced:\n- We created index assets for KB when the AI Assistant
flyout opens.\n- Even if the user does not set up the KB, they will have
a component\ntemplate pointing to the custom inference endpoint.\n\nWith
the introduction of multilingual KB:\n- We moved some of the index
creation to when setting up the KB. \n- We try to do the semantic_text
migration at startup. During this\nmigration, for users who didn't set
up the KB but had the index assets\ncreated at startup, the custom
inference endpoint will be unavailable.\n- But since the migration uses
the inference endpoint from the write\nindex, we try to access an
endpoint that's not available.\n\nThis is the reason for this error to
be logged.\n```\nInference endpoint \"obs_ai_assistant_kb_inference\"
not found or unavailable: resource_not_found_exception\n\tRoot
causes:\n\t\tresource_not_found_exception: Inference endpoint not found
[obs_ai_assistant_kb_inference]\n```\n\nThere is no customer impact from
this, just that the error that gets\nlogged is creating a lot of
noise.\n\n## Solution\n\nThis PR checks whether there are documents
without semantic_text before\nstarting the migration.\nAnd also reduced
the log level to warn because we hit the `/status`\nendpoint once when a
user opens the AI Assistant.\n\n\n### Checklist\n\n- [x] The PR
description includes the appropriate Release Notes section,\nand the
correct `release_note:*` label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"b1e7012477af991632a0ad2fdfccad743bcac3c6"}},"sourceBranch":"main","suggestedTargetBranches":["8.19"],"targetPullRequestStates":[{"branch":"main","label":"v9.1.0","branchLabelMappingKey":"^v9.1.0$","isSourceBranch":true,"state":"MERGED","url":"https://github.com/elastic/kibana/pull/221152","number":221152,"mergeCommit":{"message":"[Obs
AI Assistant] Check for documents before starting semantic text
migration (#221152)\n\nCloses
https://github.com/elastic/kibana/issues/221157\n\n## Summary\n\nWe run
a semantic text migration at startup to add the semantic text\nfield to
documents that were created before 8.17.\n\nBefore multilingual KB was
introduced:\n- We created index assets for KB when the AI Assistant
flyout opens.\n- Even if the user does not set up the KB, they will have
a component\ntemplate pointing to the custom inference endpoint.\n\nWith
the introduction of multilingual KB:\n- We moved some of the index
creation to when setting up the KB. \n- We try to do the semantic_text
migration at startup. During this\nmigration, for users who didn't set
up the KB but had the index assets\ncreated at startup, the custom
inference endpoint will be unavailable.\n- But since the migration uses
the inference endpoint from the write\nindex, we try to access an
endpoint that's not available.\n\nThis is the reason for this error to
be logged.\n```\nInference endpoint \"obs_ai_assistant_kb_inference\"
not found or unavailable: resource_not_found_exception\n\tRoot
causes:\n\t\tresource_not_found_exception: Inference endpoint not found
[obs_ai_assistant_kb_inference]\n```\n\nThere is no customer impact from
this, just that the error that gets\nlogged is creating a lot of
noise.\n\n## Solution\n\nThis PR checks whether there are documents
without semantic_text before\nstarting the migration.\nAnd also reduced
the log level to warn because we hit the `/status`\nendpoint once when a
user opens the AI Assistant.\n\n\n### Checklist\n\n- [x] The PR
description includes the appropriate Release Notes section,\nand the
correct `release_note:*` label is applied per
the\n[guidelines](https://www.elastic.co/guide/en/kibana/master/contributing.html#kibana-release-notes-process)","sha":"b1e7012477af991632a0ad2fdfccad743bcac3c6"}},{"branch":"8.19","label":"v8.19.0","branchLabelMappingKey":"^v(\\d+).(\\d+).\\d+$","isSourceBranch":false,"state":"NOT_CREATED"}]}]
BACKPORT-->

Co-authored-by: Viduni Wickramarachchi <viduni.wickramarachchi@elastic.co>
@sorenlouv sorenlouv mentioned this pull request May 22, 2025
Merged
akowalska622 pushed a commit to akowalska622/kibana that referenced this pull request May 29, 2025
@viduni94 @akowalska622
[Obs AI Assistant] Check for documents before starting semantic text …
6017f14 
…migration (elastic#221152)

Closes elastic#221157

## Summary

We run a semantic text migration at startup to add the semantic text
field to documents that were created before 8.17.

Before multilingual KB was introduced:
- We created index assets for KB when the AI Assistant flyout opens.
- Even if the user does not set up the KB, they will have a component
template pointing to the custom inference endpoint.

With the introduction of multilingual KB:
- We moved some of the index creation to when setting up the KB. 
- We try to do the semantic_text migration at startup. During this
migration, for users who didn't set up the KB but had the index assets
created at startup, the custom inference endpoint will be unavailable.
- But since the migration uses the inference endpoint from the write
index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.
```
Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]
```

There is no customer impact from this, just that the error that gets
logged is creating a lot of noise.

## Solution

This PR checks whether there are documents without semantic_text before
starting the migration.
And also reduced the log level to warn because we hit the `/status`
endpoint once when a user opens the AI Assistant.


### Checklist

- [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)
viduni94 added a commit to viduni94/kibana that referenced this pull request Jun 11, 2025
@viduni94
[Obs AI Assistant] Check for documents before starting semantic text …
4302bc7 
…migration (elastic#221152)

Closes elastic#221157

## Summary

We run a semantic text migration at startup to add the semantic text
field to documents that were created before 8.17.

Before multilingual KB was introduced:
- We created index assets for KB when the AI Assistant flyout opens.
- Even if the user does not set up the KB, they will have a component
template pointing to the custom inference endpoint.

With the introduction of multilingual KB:
- We moved some of the index creation to when setting up the KB.
- We try to do the semantic_text migration at startup. During this
migration, for users who didn't set up the KB but had the index assets
created at startup, the custom inference endpoint will be unavailable.
- But since the migration uses the inference endpoint from the write
index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.
```
Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]
```

There is no customer impact from this, just that the error that gets
logged is creating a lot of noise.

## Solution

This PR checks whether there are documents without semantic_text before
starting the migration.
And also reduced the log level to warn because we hit the `/status`
endpoint once when a user opens the AI Assistant.

### Checklist

- [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)

(cherry picked from commit b1e7012)

# Conflicts:
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/inference_endpoint.ts
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/startup_migrations/populate_missing_semantic_text_fields.ts
@viduni94 viduni94 mentioned this pull request Jun 11, 2025
Closed
@viduni94 viduni94 removed backport:version Backport to applied version labels v9.0.3 labels Jun 12, 2025
@viduni94 viduni94 added backport:version Backport to applied version labels v9.0.3 v8.18.3 and removed v8.18.3 labels Jun 12, 2025
@kibanamachine
Copy link
Contributor

kibanamachine commented Jun 12, 2025

Starting backport for target branches: 8.18, 8.19, 9.0

https://github.com/elastic/kibana/actions/runs/15609348954

@kibanamachine
Copy link
Contributor

kibanamachine commented Jun 12, 2025

💔 All backports failed

Status Branch Result
❌ 8.18 Backport failed because of merge conflicts
❌ 8.19 Cherrypick failed because the selected commit (b1e7012) is empty. It looks like the commit was already backported in #221203
❌ 9.0 Backport failed because of merge conflicts

Manual backport

To create the backport manually run:

node scripts/backport --pr 221152

Questions ?

Please refer to the Backport tool documentation

viduni94 added a commit to viduni94/kibana that referenced this pull request Jun 12, 2025
@viduni94
[Obs AI Assistant] Check for documents before starting semantic text …
cb2ce8a 
…migration (elastic#221152)

Closes elastic#221157

## Summary

We run a semantic text migration at startup to add the semantic text
field to documents that were created before 8.17.

Before multilingual KB was introduced:
- We created index assets for KB when the AI Assistant flyout opens.
- Even if the user does not set up the KB, they will have a component
template pointing to the custom inference endpoint.

With the introduction of multilingual KB:
- We moved some of the index creation to when setting up the KB.
- We try to do the semantic_text migration at startup. During this
migration, for users who didn't set up the KB but had the index assets
created at startup, the custom inference endpoint will be unavailable.
- But since the migration uses the inference endpoint from the write
index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.
```
Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]
```

There is no customer impact from this, just that the error that gets
logged is creating a lot of noise.

## Solution

This PR checks whether there are documents without semantic_text before
starting the migration.
And also reduced the log level to warn because we hit the `/status`
endpoint once when a user opens the AI Assistant.

### Checklist

- [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)

(cherry picked from commit b1e7012)

# Conflicts:
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/inference_endpoint.ts
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/startup_migrations/populate_missing_semantic_text_fields.ts
@viduni94 viduni94 mentioned this pull request Jun 12, 2025
Closed
viduni94 added a commit to viduni94/kibana that referenced this pull request Jun 12, 2025
@viduni94
[Obs AI Assistant] Check for documents before starting semantic text …
190f4f1 
…migration (elastic#221152)

Closes elastic#221157

## Summary

We run a semantic text migration at startup to add the semantic text
field to documents that were created before 8.17.

Before multilingual KB was introduced:
- We created index assets for KB when the AI Assistant flyout opens.
- Even if the user does not set up the KB, they will have a component
template pointing to the custom inference endpoint.

With the introduction of multilingual KB:
- We moved some of the index creation to when setting up the KB.
- We try to do the semantic_text migration at startup. During this
migration, for users who didn't set up the KB but had the index assets
created at startup, the custom inference endpoint will be unavailable.
- But since the migration uses the inference endpoint from the write
index, we try to access an endpoint that's not available.

This is the reason for this error to be logged.
```
Inference endpoint "obs_ai_assistant_kb_inference" not found or unavailable: resource_not_found_exception
	Root causes:
		resource_not_found_exception: Inference endpoint not found [obs_ai_assistant_kb_inference]
```

There is no customer impact from this, just that the error that gets
logged is creating a lot of noise.

## Solution

This PR checks whether there are documents without semantic_text before
starting the migration.
And also reduced the log level to warn because we hit the `/status`
endpoint once when a user opens the AI Assistant.

### Checklist

- [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)

(cherry picked from commit b1e7012)

# Conflicts:
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/inference_endpoint.ts
#	x-pack/platform/plugins/shared/observability_ai_assistant/server/service/startup_migrations/populate_missing_semantic_text_fields.ts
@viduni94 viduni94 mentioned this pull request Jun 12, 2025
Closed
@viduni94
Copy link
Contributor Author

viduni94 commented Jun 12, 2025

💚 All backports created successfully

Status Branch Result
✅ 9.0
✅ 8.18

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

Questions ?

Please refer to the Backport tool documentation

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@sorenlouv sorenlouv sorenlouv approved these changes

Assignees

@viduni94 viduni94

Labels

backport:version Backport to applied version labels ci:project-deploy-observability Create an Observability project release_note:fix Team:Obs AI Assistant Observability AI Assistant v8.19.0 v9.1.0

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

[Obs AI Assistant] Update semantic_text migration to avoid referencing an unavailable inference endpoint

4 participants

@viduni94 @elasticmachine @kibanamachine @sorenlouv