Skip to content

Limits

Data model

LimitValue
Namespace name[a-z0-9][a-z0-9_-]{0,62} (1–63 characters)
Relationships between two nodesOne per (type, start node, end node); a second CREATE of the same triple replaces the first
Node idA UUID, assigned at creation (or supplied as _id by the bulk API)
Writers per namespaceOne at a time; any number of readers
Constraints and indexesOn node properties only
Integer values64-bit signed; arithmetic wraps on overflow
Temporal typesDATE and UTC DATETIME only; no durations, times, zones or points

Per-node property limits

Checked when node data files are built (flush and compaction). Exceeding one fails the build with an explicit error rather than truncating data.

LimitDefaultVariable
Distinct property names per data file65,536NAMIDB_NODE_PROPERTY_MAX_PROPERTIES
Property name length1,024 bytesNAMIDB_NODE_PROPERTY_MAX_NAME_BYTES
One property value64 MiBNAMIDB_NODE_PROPERTY_MAX_VALUE_BYTES
Nesting depth of a value32NAMIDB_NODE_PROPERTY_MAX_VALUE_DEPTH
Elements in one value1,000,000NAMIDB_NODE_PROPERTY_MAX_VALUE_ELEMENTS
Decoded property page65 MiBNAMIDB_NODE_PROPERTY_MAX_PAGE_BYTES
Property catalog per file16 MiBNAMIDB_NODE_PROPERTY_MAX_CATALOG_BYTES

Query budgets (server)

LimitDefaultSettingResult when exceeded
Read statement time30 s--query-timeout504 timeout
Write statement time60 s--write-timeout504 timeout, nothing committed
Rows any read operator may materialise1,000,000--query-row-cap413 row_cap
Working memory of one statement50% of --memory-max-bytes (off without a ceiling)--query-memory-max-bytes413 query_memory_limit
Working memory of all running statements75% of --memory-max-bytesderived503 memory_pressure
Whole HTTP request≥ 120 s (derived)--http-request-timeout408
Wait for the writer lock30 s--writer-lock-timeout503
Wait for read admission5 sNAMIDB_QUERY_ADMISSION_WAIT_MS503
Concurrent readsCPU cores (at least 4)NAMIDB_MAX_CONCURRENT_QUERIESqueued, then 503
Concurrent full scans4NAMIDB_MAX_CONCURRENT_SCANSqueued
Concurrent HTTP requests1,024NAMIDB_HTTP_MAX_CONCURRENCYqueued

Results are materialised before they are returned, which is why the row cap and the per-query memory budget exist: they stop a runaway cross product or a huge UNWIND before it exhausts memory. Page large results with ORDER BY … SKIP … LIMIT or a key range. See Memory.

Every operator checks the deadline and the cancel flag as it works (since 3.3.0, cross products, joins, filters, projections, DISTINCT, grouping, sorting and list expressions included), so a statement stops within milliseconds of its budget or of POST /v0/admin/queries/:id/cancel. Avoid unanchored cartesian products of large labels (MATCH (a:Big), (b:Big)) anyway: they are expensive and usually hit the row cap or the timeout.

The embedded Python client has no default budgets; pass timeout_ms and max_rows per call.

Request and message sizes

LimitDefaultSetting
HTTP /v0/cypher body2 MiBfixed
Bolt message before authentication64 KiBfixed
Bolt message after authentication64 MiB--bolt-max-message-bytes
Bolt connections1,024NAMIDB_BOLT_MAX_CONNECTIONS
Bolt transaction idle time30 s--bolt-tx-timeout
Bolt transaction lifetime5 minNAMIDB_BOLT_MAX_TX_LIFETIME

Send large writes as several UNWIND $rows batches within these sizes.

LimitDefaultSetting
Unbounded full-text result size64 MiBNAMIDB_SEARCH_MAX_RESULT_BYTES (413 search_result_limit)
Document size for exact BM251 MiBNAMIDB_BM25_MAX_DOCUMENT_BYTES (413 search_document_limit)
Transient search memory256 MiBNAMIDB_SEARCH_WORKSPACE_MAX_BYTES (503 search_workspace_capacity)
Distinct string/boolean values per filter property inside a vector index4,096fixed; above it the property is filtered by the slower exact path
Filter postings per vector index body64 MiBfixed; same fallback
Sidecar candidates for a filter the index cannot apply8,192NAMIDB_VECTOR_FILTER_ID_CANDIDATE_CAP (0 disables)
Filtered hybrid BM25 candidates65,536NAMIDB_HYBRID_TEXT_FILTER_CANDIDATE_CAP
Live segments per search index8 (2–32)NAMIDB_SEARCH_LSM_MAX_SEGMENTS
int8 quantizationMETRIC cosine only—

Hitting a filter or candidate cap changes cost, not results: the query falls back to an exact path.

Multi-tenant capacity

LimitDefaultSetting
Open namespaces100 (oldest idle evicted)--max-namespaces (0 = unlimited)
Idle time before eviction1 h--namespace-idle-timeout
Concurrent reads per namespaceoffNAMIDB_MAX_CONCURRENT_QUERIES_PER_NAMESPACE
Concurrent backups / restores per process1 (waiters time out after 30 s)fixed

Throughput

All writes to a namespace go through its single writer, and each commit is a round trip to the store. On an object store, interactive write throughput is therefore bounded by commit latency; batch writes with UNWIND, use the Python bulk API, or enable --group-commit-window to share commits between concurrent writers. Reads scale with cores and are not blocked by writes. Spread independent graphs over namespaces to scale writes further.