---
name: arangodb
description: Expert ArangoDB multi-model database assistance covering AQL, document collections, graphs, and Foxx microservices. Use when querying graph relationships, joining document stores, or designing multi-model schemas.
---

# ArangoDB

ArangoDB is a native multi-model database. It allows you to store data as Key/Values, JSON Documents, and Graphs, and query them all with a single language (AQL).

## When to Use

- **Multi-Model Data Architectures**: Unifying document (JSON), graph, and key-value models into a single query engine with ACID transactions.
- **Complex Graph Traversals**: Executing high-performance depth-first and breadth-first traversals with edge attributes in AQL.
- **Unified Knowledge Graphs**: Storing entity metadata as rich documents while querying interconnected relationships as graphs.
- **Search & Full-Text Integration**: Leveraging ArangoSearch for integrated full-text indexing, BM25 scoring, and geo-spatial queries.

## Quick Start

```javascript
// AQL (ArangoDB Query Language) - SQL-like
FOR u IN users
  FILTER u.active == true
  FOR order IN OUTBOUND u orders
    RETURN { user: u.name, order: order.product }
```

## Core Concepts

### Multi-Model Document & Edge Collections

Documents store unstructured JSON entities; Edge collections define directed connections between documents using `_from` and `_to` attributes:

```json
// Vertex Document (users/alice)
{ "_key": "alice", "name": "Alice Mercer", "role": "Lead Architect" }

// Edge Document (knows/edge_1)
{ "_from": "users/alice", "_to": "users/bob", "relationship": "mentor", "since": 2024 }
```

### ArangoDB Query Language (AQL) Graph Traversal

Traverses interconnected relationship graphs up to N hops:

```aql
// 1 to 3 hops outward traversal from Alice
FOR v, e, p IN 1..3 OUTBOUND 'users/alice' GRAPH 'socialGraph'
  FILTER e.relationship == 'mentor'
  RETURN {
    colleague: v.name,
    distance: LENGTH(p.edges),
    path: p.vertices[*].name
  }
```

### Distributed Cluster Sharding (OneShard & SmartGraphs)

Ensures interconnected graph vertices and edges are colocated on identical physical DB servers to eliminate network hop penalties:

```javascript
// Creates a SmartGraph partitioned by customer tenant ID
const graph = db._createSmartGraph(
  "tenantSocialGraph",
  [{ collection: "members", from: ["users"], to: ["users"] }],
  { smartGraphAttribute: "tenantId" },
);
```

## Common Patterns

### Graph Traversal with AQL

**Problem**: Querying deeply nested social or hierarchy connections in relational stores requires slow multi-table recursive joins.

**Solution**:
Traverse nodes and edges declaratively with AQL:

```aql
FOR v, e, p IN 1..3 OUTBOUND 'users/alice' GRAPH 'social_graph'
  FILTER v.status == 'active'
  RETURN {
    friend: v.name,
    distance: LENGTH(p.edges),
    path: p.vertices[*].name
  }
```

## Best Practices

**Do**:

- Leverage SmartGraphs for Multi-Tenant Clusters: Colocate graph subtrees on single cluster nodes using `smartGraphAttribute` to eliminate network hop latencies.
- Use ArangoSearch Views: Replace separate external Elasticsearch clusters by indexing collections directly with ArangoSearch.
- Filter Early in Graph Traversals: Use `PRUNE` and inline `FILTER` conditions to stop traversals down dead-end relationship paths.
- Bind Parameters in AQL: Always pass variables as `@param` to prevent AQL injection and maximize query compilation cache reuse.

**Don't**:

- Perform unbounded graph traversals: Never run `1..99` traversals without a `PRUNE` condition or depth limits.
- Use Edge collections without indexes: ArangoDB automatically indexes `_from` and `_to`; avoid overriding them with inefficient manual indexes.
- Mix cross-tenant data without SmartGraph keys: Sharding graphs randomly across nodes causes severe distributed RPC overhead.

## Troubleshooting

| Error                              | Cause                                                    | Solution                                                                             |
| :--------------------------------- | :------------------------------------------------------- | :----------------------------------------------------------------------------------- |
| `[1200] conflict`                  | Optimistic locking collision on document update.         | Retry update using current document `_rev` or use `overwriteMode: 'replace'`.        |
| `[1501] collection not found`      | Attempted query against non-existent collection or typo. | Verify collection names and ensure creation migrations have executed.                |
| `Out of memory in graph traversal` | Unbounded traversal depths on large dense graphs.        | Specify explicit min..max depths and add prune/filter conditions early in traversal. |

## References

- [ArangoDB Documentation](https://docs.arangodb.com/)


---
Source attribution: G1Joshi
Original project: https://github.com/G1Joshi/Agent-Skills/blob/6a74cf03a8a86dc7d993ead3091b41123dbd0181/skills/databases/arangodb/SKILL.md
License: MIT
License information: https://opensource.org/license/mit
Original upstream instructions; downloaded from Dave AI.

## Upstream license notice

MIT License

Copyright (c) 2026 G1Joshi

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.



# Dave AI resource metadata

Creator / publisher: G1Joshi
Original project: https://github.com/G1Joshi/Agent-Skills/blob/6a74cf03a8a86dc7d993ead3091b41123dbd0181/skills/databases/arangodb/SKILL.md
Attribution basis: Source repository record

Fields of work: Data & Analytics, Development & Engineering
Roles: Backend Developer, Data Engineer, Data Scientist, Database Administrator, ML / AI Engineer
Specialties: API, Database
Tools: Arangodb
Country applicability: Country not specified
Classification basis: Source catalog labels
Instruction SHA-256: 17d4c048a296aacff1bfc136ce5df13a6a3cd39c5d21fc47e39c0f6be0c4dc77