home shape


Subject to change any time.


right blob img min

ArangoDB 3.12 - released on March ??? 2024 ✓

Improved memory accounting

Version 3.12 features multiple improvements to the observability of ArangoDB deployments. Memory usage is more accurately tracked and additional metrics have been added for monitoring the memory consumption. Furthermore, the memory usage of some subsystems has been optimized.

New ArangoSearch features

The so-called WAND optimization lets you retrieve search results for the highest-ranking matches from Views faster for a list of sort expressions that you define upfront.

The new wildcard Analyzer in combination with an inverted index or View can accelerate LIKE searches with _ and % wildcards. This is especially useful if you want to find non-prefix partial matches in long strings.

The new multi_delimiter Analyzer simplifies splitting up text using multiple delimiters.

AQL improvements

The UPSERT operation accepts a FILTER expression to select a document using complex conditions instead of matching by example.

The AQL optimizer automatically recognizes whether a merge-join strategy can be used if two or more collections are joined using nested FOR loops and the attributes you join on are indexed by primary indexes or persistent indexes.

Certain operations of a query asynchronously prefetch the next batch of data while processing the current batch, allowing parts of the query to run in parallel for better performance. Additionally, the filtering of lists and index-covered attributes has been improved.

Multi-dimensional indexes

Multi-dimensional indexes allow you to index two- or higher-dimensional data such as time ranges, for efficient intersection of multiple range queries. Version 3.12 supports two variants, one with only the dimensional data, and another to specify additional document attributes for narrowing down the search space using equality checks. The latter can be used as a vertex-centric index for AQL graph traversals.

Request and response compression

Transparent compression of requests and responses between ArangoDB servers and client tools helps to speed up data transfers and reduce traffic.

Faster dumping and restoring

Creating logical backups with arangodump can be several times faster than before thanks to parallel dumping at the shard level, as well as transfer compression and file splitting. It also improves the performance of restoring dumps.

Upgraded V8

Version 3.12 ships with the V8 JavaScript engine in version 12.1, bringing various language features to JavaScript contexts in ArangoDB like arangosh and Foxx. These features are part of the ECMAScript specifications ES2020 through ES2024, for example, optional chaining, the nullish coalescing operator, Object.groupBy(), and extended internationalization APIs.

ArangoDB 3.11 - released on May 30th 2023 ✓

Faster query performance across search and graph

ArangoDB 3.11 brings new functionality to ArangoSearch that improves its performance. New column caching options allow for common values to be stored in memory, and a new Analyzer for GeoJSON data stores geo-spatial data more efficiently. In addition, reduced memory usage of in-memory edge indexes as well as index cache refilling accelerate graph traversal performance while consuming fewer resources.

Data science and analytics operational enhancements

ArangoDB 3.11 offers persisted Pregel execution statistics to better assist data science and graph analytics use cases.
Further features in ArangoDB 3.11 include parallel gather, write-write conflict improvements, various performance enhancements across AQL (the ArangoDB Query Language), new server options, and more.

Packaging and availability

ArangoDB 3.11 is available immediately. All new features are included in both the open source ArangoDB Community Edition and the commercial ArangoDB Enterprise Edition, except for column caching and the geo_s2 Analyzer, which are Enterprise only.

Improved user experience for database administration and management

With new features available in its web interface, even more of ArangoDB’s capabilities can be managed outside of the command line in a user-friendly graphical interface. The graph viewer for visualizing named graphs has received substantial improvements as well as new functionality. It has been reimplemented based on the vis.js library, the interface has been streamlined and rewritten to use the React framework, and overall performance has been improved. New features include the ability to search the visible graph to center a specific vertex, as well as by Start node. New functionality for ArangoSearch has also been added to ArangoDB’s web interface, including a search-alias View to create and manage search views, the ability to create inverted indexes on collections, as well as a new sorting mechanism and search box for Saved Queries to make it simpler to find relevant queries and discover which ones have been created or used recently.

ArangoDB 3.11 is available for download as a self-managed deployment here, as well as on ArangoGraph Insights Platform, ArangoDB’s managed service.

Learn more

To get started with ArangoDB in a few clicks: Sign up for a free trial of ArangoGraph Insights Platform

scroll down line
right blob min

Past releases
ArangoDB 3.10 - released on October 4th 2022✓

Native ARM support

Packages for the ARM architecture, including native support for Apple silicon.

All Shortest Paths

A graph traversal algorithm to query for all paths of shortest length between two documents.

Computed Values

Persistent document attributes that are generated when documents are created or modified, using an AQL expression.

EnterpriseGraphs (EE)

A specialized version of SmartGraphs, with an automatic sharding key selection.

Parallelism for sharded graphs (EE)

Parallel execution of AQL traversal queries with many start vertices for all types of sharded graphs, leading to faster results.

Traversal projections (EE)

Optimized data loading for AQL traversal queries if only a few document attributes are accessed.

Persistent indexes

– Optional in-memory cache for faster lookups
arded graphs, leading to faster results.
– Stored values to let persistent indexes cover additional attributes of projections.

Inverted indexes and lightweight Views

An eventually consistent index type that can accelerate a broad range of queries, defined per collection, that can be added to search-alias Views for searching multiple collections at once and for ranking capabilities.

ArangoSearch MinHash (EE)

Locality-sensitive hashing with index support to approximate the Jaccard similarity for entity resolution.

Search highlighting (EE)

Get the substring positions of matched terms, phrases, or n-grams.

Nested search (EE)

Match arrays of objects with all the conditions met by a single sub-object, and define for how many of the elements this must be true.

scroll down line
right blob img min

ArangoDB 3.9 - released on February 15th 2022 ✓

SmartGraphs using SatelliteCollections (EE)

SatelliteCollections can be used in (Disjoint) SmartGraphs to enable more local execution of graph queries.


– Segmentation Analyzer for language-agnostic tokenization of text
– Collation Analyzer to honor a language’s alphabetical order in range queries

scroll down line

ArangoDB 3.8 - released on July 29th 2021 ✓

k Paths and Weighted Traversals

Enumerate all paths between two vertices. Emit paths in order of increasing weights.

Sliding Window Queries

AQL WINDOW operation to aggregate adjacent documents, value ranges and time intervals.


– Geo-spatial queries
– Pipeline Analyzer to combine the effects of multiple Analyzers into one
– AQL Analyzer to pre-process fields and to filter out documents
– Stopwords Analyzer to discard tokens

Encryption at Rest (EE)

AES hardware acceleration support.

scroll down line

ArangoDB 3.7 - released on September 16th 2020 ✓


Graphs that will be replicated on all nodes of a cluster to speed-up queries that join local data.

Schema Validation

Specify JSON schema (draft-4) validation to enforce schema during updates and inserts in ArangoDB.


– Wildcard search
– N-Gram based fuzzy search
– Enhanced phrase and proximity search


– SuperUser JWT token rotation
– TLS certificate rotation
– Encryption Key rotation

right blob min
scroll down line

ArangoDB 3.6 - released on January 8th 2020 ✓

Parallelization of Cluster AQL queries

Run AQL queries on several DB-Servers in parallel.

Late document materialization (RocksDB)

Only fetch the relevant documents from SORT/LIMIT queries.

Subquery performance optimization

Subquery splicing inlines the execution of certain subqueries to improve execution time.

Early pruning of non-matching documents

Discard documents that don’t match the FILTER condition in full collection scans.

OneShard (EE)

Single server performance & ACID semantics with cluster fault-tolerance for single databases and also eligible for selected collections.


– Added UTF-8 support and ability to mark beginning/end of the sequence.

– Added edge n-gram support in Text Analyzer type

Dynamic search expressions with arrays

– Support for array comparison operators
– Enhanced usability with array support in TOKENS/PHRASE functions

scroll down line

ArangoDB 3.5 - released on August 21st 2019 ✓



Streaming Transaction API

Perform a multi-document transaction with individual begin and commit / abort commands.

Configurable Analyzers

Create custom text processors using new and extended types of Analyzers.

Time-to-Live Index (TTL)

Remove expired documents from a collection.

Sorted Views

Define a primary sort order upfront to speedup View sorting tremendously.

AQL Improvements

Performance improvements, parallel queries, PRUNE keyword and k Shortest Paths traversal support.

Improved AQL Integration

Return ranking scores, restrict search to collection subsets, use range searches.

SmartJoins (EE)

– Added UTF-8 support and ability to mark beginning/end of the sequence.

– Added edge n-gram support in Text Analyzer type

Data Masking

Use data masking capabilities when exporting attributes containing sensitive data / PII via arangodump.

Consistent Snapshot Backup (EE)

Consistent snapshot backup in cluster mode.

scroll down line
one query background

ArangoDB 3.4 - released on December 6th 2018 ✓


A sophisticated, integrated full-text search solution over a user-defined set of attributes and collections.

Improved geo functionality

The new geo index functionality allows indexing complex geographical objects in addition to indexing simple point coordinates. Functionality has been added for querying and comparing GeoJSON objects. Geo index performance has been improved vastly for the RocksDB engine.

Repsert operation

Insert operations can now be turned into a replace automatically, in case that the target document already exists. Such operations (called a “Repsert”) can simplify client application development.

Optimized binary format for the RocksDB engine

ArangoDB 3.4 can use an optimized binary format for storing documents with the RocksDB storage engine, allowing for better long-term insertion performance.

Round-robin load-balancer support

ArangoDB now supports running multiple coordinators behind round-robin load balancers, such as they can be found in cloud environments often.

Faster cluster AQL execution

The cluster-internal protocol for running AQL queries has been improved so that AQL queries can run in a cluster with less overhead.

AQL query profiling

AQL queries can now be profiled in detail, so that query execution plans show detailed runtime information.

Distributed COLLECT

In a cluster setup, COLLECT queries for grouping and aggregation can now execute significant parts of the query on the database servers, greatly reducing the amount of data to be transferred between database servers and the coordinator.

Performance improvements

All built-in AQL functions now have native implementations in C++, and will not fall back to using the V8 JavaScript engine. The same is true for other core APIs. This provides better performance and reduced resource usage.

Improved sparse index support

The AQL query optimizer can now use sparse indexes in more cases than it was able to in previous versions, making sparse indexes a viable option in more situations and queries.

Parallel dump and restore

ArangoDB’s tools for database backups are now multi-threaded, which means taking and restoring backups is now faster than in previous versions.

scroll down line
right blob min

ArangoDB 3.3 - released on December 22nd 2017 ✓

DC2DC Replication (EE)

Additional storage engine for ArangoDB to work with huge datasets. Document level locking on writes, no locking on reads

Encrypted backup (EE)

The encryption key can be read from a file or from a generator program. It works in single server and cluster mode.

Resilient active/passive mode

There is now a mode to start two arangod instances as a pair of connected servers with automatic failover.

Server-level replication

The new globalApplier has the same interface as the existing applier, but it will replicate from all database on the leader and not just a single one.

RocksDB throttling

It throttles write operations to RocksDB in the RocksDB storage engine, in order to prevent total stalls.

Faster shard creation in cluster

Creating collections is what all ArangoDB users do. So it should be as quick as possible.

scroll down line

ArangoDB 3.2 - released on July 20th 2017 ✓

RocksDB & Pluggable Storage Engine

Additional storage engine for ArangoDB to work with huge datasets. Document level locking on writes, no locking on reads

Distributed Graph Processing with Pregel

Use incremental graph processing algorithms in a single mode server or cluster

Fault-Tolerant Foxx

Foxx service are now self-healing, even if all coordinators go down.


Get documents sorted by distance to a certain point in space. You can also apply filters and limits to geo_cursor. E.g. “Give me 10 vegetarian restaurants within a 1 mile radius to X”


Export your data in multiple formats. Export graphs data to XGMML format for Cytoscape visualizations or arbitrary collections to JSON or JSONL.

Satellite Collections (EE)

Satellite Collections enable faster join operations when working with sharded datasets and avoid expensive network hops during join processing among machines.

Encryption at Rest (EE)

Even if a disk gets stolen, data can’t be accessed.


ArangoDB can now be integrated with LDAP allowing for an external authentication server to manage users.

one query background
scroll down line

ArangoDB 3.1 - released November 3rd 2016✓

VelocyPack over HTTP

Stream binary storage VelocyPack over HTTP


Directly stream our binary format VelocyPack for high performance needs

boost-ASIO server infrastructure

Performance boost with new boost-ASIO

Stand-Alone Agency

Use ArangoDB as a resilient, RAFT-based key/value store as alternative to ZooKeeper or etcd

AQL Editor

Much easier to use. Choose JSON, tabular or graph outputs. Simplified elaboration of queries with new Query Performance Profiler

New Graph Viewer

Suitable for large graph visualization with much more features. First WebGL implementation

Overhauled Query Optimizer

Improved abstraction to integrate pluggable storage engine and MVCC

Preparations for pluggable storage engine and MVCC

Suitable for large graph visualization with much more features. First WebGL implementation

Vertex-Centric Indices

Generate indices on edges which are a combination of vertex and attribute

New Java Driver

Multi-document operations, VelocyStream ready, asynchronous request handling

SmartGraphs (EE)

Shard large graph datasets to a cluster and stay close to the performance of a single instance

Auditing (EE)

Keep a detailed log of all the important things that happened in ArangoDB

Encryption Control (EE)

Choose your level of SSL encryption

scroll down line
right blob min

ArangoDB 3.0 - released on July 23rd 2016 ✓


Internal storage will change from JSON to VelocyPack for enhanced performance, smaller footprint and binary support.

Persistent Indexes

We plan on making indexes persistent, which will allow using quicker recovery, start-up and larger datasets.

Low-Level C++ Driver

Implementation of efficient, reusable, platform-independent core driver functionality to be used in multiple client languages.


Allow for automatic failover to slave nodes. A monitor process detects network failures and automatically switches to backup nodes.

Master/Master Replication

Replicate data not just in a master/slave fashion, but also as true master/master.

Automatic Failover with Mesos

This release will contain the next iteration of our Mesosphere DCOS integration and will thus offer convenient set-up of synchronous replication and full automatic failover.

Health Check Dashboard in Mesos

Enables you to see the health status of your ArangoDB cluster in Mesos dashboard.

Cluster Dashboard

Improved cluster administration will be implemented.

Distributed Failover

Instead of dedicated slave, you can use spare capacity on masters to hold the slave for other shards.

scroll down line
one query background

ArangoDB 2.8 – released on January 25th 2016 ✓

Array Indexes

Hash indexes and skiplist indexes support array values so they index individual array members.

Graph Traversals in AQL

Using AQL to traverse a graph / edge collections.

AQL Optimization

Reimplemented AQL functions in C++ for improved performance.

New Framework in Mesosphere DCOS 1.3

ArangoDB package for DCOS 1.3, enhanced replication and failover.

Automatic Deadlock Detection for Transactions

The new deadlock detection mechanism will kick in automatically when it detects operations that are mutually waiting for each other.

scroll down line

ArangoDB 2.7 - released October 9th 2015 ✓

Replication Improvements

This allows much easier synchronization of a single collection from a master to a slave server.

Throughput Enhancements

A lot is not enough. Throughput is another key requirement for a premium database. Again we pushed our throughput a big step forward with 2.7.

Improved Date Handling in AQL

AQL functions for date and time calculation and manipulation.

Index Buckets

Split primary indexes and hash indexes into multiple index buckets.

(EE) = Available in the Enterprise Edition. Everything else is available in both the Community Edition and Enterprise Edition.