spring-projects/spring-ai · error · RuntimeException

Search failed!

Error message

Search failed!

What it means

doSimilaritySearch checks the R<SearchResults> response from the Milvus client and throws a plain RuntimeException("Search failed!") when respSearch.getException() is non-null. The Milvus server rejected or failed the similarity search RPC; the original exception is the cause.

Solutions

1. Read the cause exception to get the exact Milvus status/error.
2. Confirm the query embedding dimension matches the collection's embedding field dimension.
3. Load the collection (milvusClient.loadCollection) before searching.
4. Check the search params JSON (e.g. nprobe) is valid for the configured index type.
5. Verify Milvus server connectivity and collection existence.

Example fix

// before
SearchRequest request = SearchRequest.builder().query("...").topK(10000).build();
// after
SearchRequest request = SearchRequest.builder().query("...").topK(100).build(); // keep topK within server limits

// verify collection is loaded and dimensions match before searching
String dims = embeddingModel.dimensions(); // must equal the collection's embedding field dimension
// also confirm collection exists:
milvusClient.hasCollection(HasCollectionParam.newBuilder().withCollectionName(collectionName).build());

try {
    List<Document> docs = vectorStore.similaritySearch(SearchRequest.builder().query(q).topK(5).build());
} catch (RuntimeException e) {
    logger.error("Milvus search failed: {}", e.getCause() != null ? e.getCause().getMessage() : e.getMessage(), e);
    // optionally retry once after verifying collection load state
}

When it happens

Trigger: Calling similaritySearch(SearchRequest) when the Milvus search RPC returns an error status: collection not loaded, invalid search params (nprobe/ef), wrong embedding dimension, or connectivity failure.

Common situations: Query embedding dimension differs from the collection's index dimension; topK too large; search params JSON invalid; Milvus collection dropped or not loaded after restart.

Understand the failure class

Background: Database query failed: Internal Server Error 500s wrapping SQL, Prisma, and connection failures — what to check first — this error's family across 16 libraries.

Related errors

• Deleted only entries from requested
• Failed to delete documents by filter
• Failed to delete documents by filter:
• Failed to insert:
• Not supported expression type

AI-assisted analysis of spring-projects/spring-ai@98a7beda4f (2026-09-11). Data as JSON: /api/errors/986c89b7455247cf. Report an issue: GitHub.

Appendix: source

Thrown at vector-stores/spring-ai-milvus-store/src/main/java/org/springframework/ai/vectorstore/milvus/MilvusVectorStore.java:405

			.withFloatVectors(List.of(EmbeddingUtils.toList(embedding)))
			.withVectorFieldName(this.embeddingFieldName);

		if (StringUtils.hasText(nativeFilterExpressions)) {
			searchParamBuilder.withExpr(nativeFilterExpressions);
		}

		if (StringUtils.hasText(this.partitionName)) {
			searchParamBuilder.addPartitionName(this.partitionName);
		}

		if (StringUtils.hasText(searchParamsJson)) {
			searchParamBuilder.withParams(searchParamsJson);
		}

		R<SearchResults> respSearch = this.milvusClient.search(searchParamBuilder.build());

		if (respSearch.getException() != null) {
			throw new RuntimeException("Search failed!", respSearch.getException());
		}

		SearchResultsWrapper wrapperSearch = new SearchResultsWrapper(respSearch.getData().getResults());

		return wrapperSearch.getRowRecords(0)
			.stream()
			.filter(rowRecord -> getResultSimilarity(rowRecord) >= request.getSimilarityThreshold())
			.map(rowRecord -> {
				String docId = String.valueOf(rowRecord.get(this.idFieldName));
				String content = (String) rowRecord.get(this.contentFieldName);
				JsonObject metadata = new JsonObject();
				try {
					metadata = (JsonObject) rowRecord.get(this.metadataFieldName);
					if (metadata != null) {
						// inject the distance into the metadata.
						metadata.addProperty(DocumentMetadata.DISTANCE.value(), 1 - getResultSimilarity(rowRecord));
					}
				}

View on GitHub (pinned to 98a7beda4f)