Technologies/Meilisearch/meilisearch.search.processing_time
MeilisearchMeilisearchMetric

meilisearch.search.processing_time

Query processing duration
Dimensions:None
Available on:Native (1)
Interface Metrics (1)
Native
Time taken to process the search query in milliseconds
Dimensions:None
Sources
search.processing_timewww.meilisearch.com

Technical Annotations (25)

Configuration Parameters (6)
HANNOY_Mrecommended: slightly above 16 for 10M+ documents
controls HNSW graph connectivity and depth
searchableAttributes
exclude unique IDs and URLs to reduce database size
stopWords
essential for reducing database size
filterableAttributes
minimize to reduce disk space consumption
rankingRules
proximity rule can be 80% of database size
--experimental-drop-search-afterrecommended: match reverse proxy timeout (e.g., 10 seconds)
prevents wasting resources on searches dropped by reverse proxy
Error Signatures (1)
Idle timeout reached for "http://127.0.0.1:7700/indexes/*/search"error code
CLI Commands (1)
--experimental-drop-search-after <seconds>remediation
Technical References (17)
crates/milli/src/vector/embedder/composite.rsfile pathUnalignedVectorcomponentSIMDconceptSSE/AVXconceptHNSWconceptHANNOY_Mcomponentsemantic searchcomponentvector storecomponentdockercomponentmemory mapconceptvirtual memoryconceptreverse proxycomponenttokenizercomponentsoft space separatorconcept/searchcomponentsearch queuecomponentavailable_parallelismconcept
Related Insights (13)
Unaligned vector memory hinders SIMD distance calculation performanceinfo
▸
Default HANNOY_M parameter may cause deep graphs in large datasetsinfo
▸
Semantic search performance regression in Meilisearch 1.16+warning
▸
Experimental vector store improves semantic search performanceinfo
▸
Docker deployment has performance overhead versus standaloneinfo
▸
API response time exceeds 2 secondswarning
▸
Accessing hundreds of indexes in short periods degrades performancewarning
▸
Specific search queries trigger high CPU usagewarning
▸
Search performance degrades with large database to RAM ratiowarning
▸
CPU contention from parallel searches prevents timely resultscritical
▸
Search requests continue processing after reverse proxy timeoutwarning
▸
Hash character (#) in queries causes 100x+ performance degradationcritical
▸
Concurrent search requests exhaust CPU and RAM without queue limitswarning
▸