fix(search): don't fail the reindex when the embedding provider is unreachable

[mohityadav766]

Follows up #30364, which is on main and the 2.0 release branch.

What happens today

#30364 un-gated the staged chunk recreate, so every job-driven full recreate now calls beginStagedChunkRecreate(), which opens with a pre-flight embed. That pre-flight rethrows, and the exception fails the entire SearchIndexApp run — so a reindex that needs no embeddings at all is blocked by an optional AI provider being unreachable.

Why this is not a narrow case

Collate ships naturalLanguageSearch.semanticSearchEnabled=true by default, while llmConfiguration.embeddings.provider defaults to bedrock. And BedrockEmbeddingClient's constructor never calls AWS — it validates model id, dimension and region, then builds the SDK client.

So on any deployment where a region resolves (anything running on AWS) but bedrock:InvokeModel was never granted:

1. the client constructs happily,
2. initializeVectorSearchService reports success,
3. and the first full reindex dies:

User: arn:aws:sts::...:assumed-role/... is not authorized to perform:
bedrock:InvokeModel on resource: .../amazon.titan-embed-text-v2:0
(Service: BedrockRuntime, Status Code: 403)

surfacing as status='failed' within seconds with an empty failureContext.

The misconfiguration is otherwise invisible. Live indexing logs embedding errors and carries on, so the deployment looks healthy right up until someone reindexes. OSS is unaffected — semanticSearchEnabled defaults to false there, so the vector service never initialises and the gate returns early.

This is what has been failing the nightly Java IT reindex suites on main and 2.0 since 2026-08-04.

The change

Treat an unavailable provider as "do not stage" rather than "fail": return null, which the caller already handles as the partial-recreate outcome — existing chunks stay live and get swept by the next recreate that runs with a working provider. markEntityTypeReindexed already ignores marks from a run without staging (the "unbound chunk-type mark" branch), so this reuses a supported state rather than inventing one.

The pre-flight's intent is preserved: it still refuses to stage a generation it could never finish. It just no longer takes the entity reindex down with it.

Genuine staging failures still throw — an indeterminate live-target probe, a failed index create — because those mean continuing could destroy live chunks. The existing beginStagedChunkRecreate_abortsWhenTheLiveTargetProbeIsIndeterminate test still passes unchanged.

Test plan

• OpenSearchVectorServiceChunkStagingTest — 8/8 pass, including a new case asserting an unavailable provider returns null and touches nothing in the cluster (verify(client, never()).indices() / .generic()).
• RecreateWithEmbeddingsTest — 7/7 pass.
• Nightly Java IT once merged.

Unit tests were executed on main; the 2.0 change is character-identical (verified by diffing the two patches) but its module could not be built locally offline.

Related

• open-metadata/openmetadata-nightly#285 — points the Java IT cluster at DJL in-process embeddings, so CI stops depending on Bedrock at all. Complementary: that fixes the environment, this stops the environment from being able to break reindex.
• fix(tests): fail fast when a reindex run ends in a terminal non-success status #31126 / fix(tests): fail fast when a reindex run ends in a terminal non-success status #31127 — the harness bug that turned this failure into a 60-minute poll and cascading job timeouts.

🤖 Generated with Claude Code

[gitar-bot]

Code Review ✅ Approved

Makes embedding provider failures during search reindexing optional by returning null instead of throwing, allowing reindex jobs to complete when AI services are unreachable. No issues found.

Options

Display: compact → Showing less information.

Comment with these commands to change the behavior for this request:

Compact

gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Gitar | Powered by Gitar — free for open source}

[github-actions]

❌ PR checklist incomplete

This PR cannot be merged until the following are addressed on its linked issue:

• No GitHub issue is linked. Link an issue in the Development section of the PR (or add Fixes #12345 to the description). For a same-org cross-repo issue, add Fixes open-metadata/<repo>#123 to the description.

The fields live on the linked issue in the Shipping project (open the issue → right sidebar → Projects). After you set them, re-run this check (or push a commit) — issue/project changes do not re-trigger it automatically.

Maintainers can bypass this check by adding the skip-pr-checks label.