langchain-ai/langgraph
1
0
Fork 0
mirror of https://github.com/langchain-ai/langgraph.git synced 2026-07-15 12:35:43 -04:00
Code Issues 314 Packages Projects Releases 17 Wiki Activity

[PR #6369] feat(checkpoint): add Kusto/Azure Data Explorer/Eventhouse checkpoint implementation #4997

New Issue
Closed
opened 2026-02-20 17:51:04 -05:00 by yindo · 0 comments
yindo commented 2026-02-20 17:51:04 -05:00
Owner
Copy Link
Copy Source

Original Pull Request: https://github.com/langchain-ai/langgraph/pull/6369

State: closed
Merged: No

Adds a new checkpoint implementation for LangGraph that uses the Kusto engine (available in Fabric RTI Eventhouse, Azure Data Explorer) as the backend storage. This implementation provides enterprise-grade persistence with unique capabilities tailored for AI agent workloads.

Key Features

  • Graph + Time Series Semantics: Leverages Kusto's native graph query capabilities (KQL graph-match) and time-series analytics for traversing checkpoint relationships and analyzing agent performance over time
  • Production-Ready: Full async/sync support, streaming ingestion (<1s latency), materialized views for optimized "latest checkpoint" queries
  • Enterprise Scale: Handles millions of checkpoints with Kusto's distributed architecture
  • Multiple Deployment Options: Works with Azure Data Explorer clusters, Microsoft Fabric Eventhouse, or free clusters for testing
  • Full LangGraph Compatibility: Implements all BaseCheckpointSaver methods including human-in-the-loop support

Why Kusto?

Agent workflows are inherently both graphs (parent-child checkpoint relationships, multi-agent interactions) and time series (performance tracking, state evolution). Kusto uniquely handles both:

  • Query parent-child checkpoint relationships using graph semantics
  • Analyze agent performance trends over time
  • Track state evolution across distributed multi-agent systems
  • Full audit trail with enterprise compliance features

What's Included:

  • Complete async implementation with sync wrappers
  • JSON serialization for checkpoint data (stored in Kusto's dynamic columns)
  • Comprehensive documentation (README, SETUP, QUICKSTART guides)
  • 5 tutorials covering: Basic usage and first checkpoint
    , simple agent with conversation memory
    , production-ready patterns with error handling
    , OpenAI/Azure OpenAI chatbot integration
    , multi-agent collaboration with parallel execution
    , Kusto table provisioning script (provision.kql)

Issue:

N/A (new feature)

Dependencies:

All dependencies are standard Azure SDK packages widely used in production:

  • azure-kusto-data[aio]>=4.3.1 - Query client with async support
  • azure-kusto-ingest>=4.3.1- Streaming ingestion client
  • azure-identity>=1.15.0 - Azure authentication
  • aiohttp>=3.8.0 - Required by azure-kusto-data async
  • orjson>=3.10.1 - Fast JSON serialization

Note on Testing:

The implementation follows the BaseCheckpointSaver contract exactly. Validation is provided through comprehensive tutorials in "examples" directory that demonstrate real-world usage patterns. The make test target passes but acknowledges this approach rather than traditional unit tests.

Twitter handle: @cosh23

**Original Pull Request:** https://github.com/langchain-ai/langgraph/pull/6369 **State:** closed **Merged:** No --- Adds a new checkpoint implementation for LangGraph that uses the Kusto engine (available in Fabric RTI Eventhouse, Azure Data Explorer) as the backend storage. This implementation provides enterprise-grade persistence with unique capabilities tailored for AI agent workloads. ### Key Features - **Graph + Time Series Semantics**: Leverages Kusto's native graph query capabilities (KQL graph-match) and time-series analytics for traversing checkpoint relationships and analyzing agent performance over time - **Production-Ready**: Full async/sync support, streaming ingestion (<1s latency), materialized views for optimized "latest checkpoint" queries - **Enterprise Scale**: Handles millions of checkpoints with Kusto's distributed architecture - **Multiple Deployment Options:** Works with Azure Data Explorer clusters, Microsoft Fabric Eventhouse, or free clusters for testing - F**ull LangGraph Compatibility**: Implements all BaseCheckpointSaver methods including human-in-the-loop support ### Why Kusto? Agent workflows are inherently both graphs (parent-child checkpoint relationships, multi-agent interactions) and time series (performance tracking, state evolution). Kusto uniquely handles both: - Query parent-child checkpoint relationships using graph semantics - Analyze agent performance trends over time - Track state evolution across distributed multi-agent systems - Full audit trail with enterprise compliance features ### What's Included: - Complete async implementation with sync wrappers - JSON serialization for checkpoint data (stored in Kusto's dynamic columns) - Comprehensive documentation (README, SETUP, QUICKSTART guides) - 5 tutorials covering: Basic usage and first checkpoint , simple agent with conversation memory , production-ready patterns with error handling , OpenAI/Azure OpenAI chatbot integration , multi-agent collaboration with parallel execution , Kusto table provisioning script (provision.kql) ### Issue: N/A (new feature) ### Dependencies: All dependencies are standard Azure SDK packages widely used in production: - azure-kusto-data[aio]>=4.3.1 - Query client with async support - azure-kusto-ingest>=4.3.1- Streaming ingestion client - azure-identity>=1.15.0 - Azure authentication - aiohttp>=3.8.0 - Required by azure-kusto-data async - orjson>=3.10.1 - Fast JSON serialization ### Note on Testing: The implementation follows the BaseCheckpointSaver contract exactly. Validation is provided through comprehensive tutorials in "examples" directory that demonstrate real-world usage patterns. The make test target passes but acknowledges this approach rather than traditional unit tests. Twitter handle: @cosh23
yindo added the pull-request label 2026-02-20 17:51:04 -05:00
yindo closed this issue 2026-02-20 17:51:04 -05:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agents
awaiting author revision
bug
change
chat models
checkpointer
ci
code quality
code quality
dependencies
documentation
drawing
duplicate
enhancement
good first issue
help wanted
invalid
maintainer
meta
missing information
not maintained by langchain
open-swe-auto
open-swe-max
open-swe-max-auto
pending
performance
pull-request
Mirrored from GitHub Pull Request
 question
requires design
sdk
server
stale
studio
tests
type checking
usage question
v0.5
v1.1
No Label pull-request
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
yindo
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: langchain-ai/langgraph#4997
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?