Issues

• Reclaim failed SnapshotJob non-target artifacts when the source Pod record is gone

  enhancement
  ai-dynamo/snapshot#695 · Filter by author hhzhang16 opened on Oct 22, 2026

• Preserve TCP connections between set members

  enhancement
  Task

  ai-dynamo/snapshot#693 · Filter by author julienmancuso opened on Oct 22, 2026
  

• Serve NCCL Checkpoint's rendezvous through the agent

  enhancement
  Task

  ai-dynamo/snapshot#692 · Filter by author julienmancuso opened on Oct 22, 2026
  

• [Impl] Support Fallback to coldstart on restore failures

  ai-dynamo/snapshot#689 · Filter by author danbar2 opened on Oct 22, 2026 · v0.3.0
  

• rename checkpointTimeoutSeconds and restoreTimeoutSeconds

  bug
  Bug

  ai-dynamo/snapshot#688 · Filter by author danbar2 opened on Oct 22, 2026
  

• Deleted image files come back after restore because kernel overlayfs whiteouts are not recognized

  ai-dynamo/snapshot#676 · Filter by author v0ropaev opened on Oct 21, 2026

• ApplyDeletedFiles skips every entry when the target root is "/"

  ai-dynamo/snapshot#675 · Filter by author v0ropaev opened on Oct 21, 2026

• vLLM DEP8 (DeepGEMM Mega MoE) deadlocks in decode under cuInterpose with UMD 615 on B300; on UMD 580 it runs but cannot restore (#510)

  approved
  Bug

  ai-dynamo/snapshot#666 · Filter by author kbespalov opened on Oct 20, 2026
  2

  

• Expose Prometheus metrics for capture and restore (operator and agent)

  enhancement
  Feature

  ai-dynamo/snapshot#665 · Filter by author ozbarshalom opened on Oct 20, 2026

• Verify checkpoint reuse through scaling and replacement, with matched first-response benchmarks

  approvedgood first issue
  Task

  ai-dynamo/snapshot#655 · Filter by author michaelshin opened on Oct 19, 2026
  

• [DOC]: Add an Argo CD capture-to-restore integration guide

  approveddocumentationgood first issue
  Task

  ai-dynamo/snapshot#654 · Filter by author michaelshin opened on Oct 19, 2026
  1

  

• [DOC]: Architecture reference is an empty placeholder (docs/reference/architecture.md)

  approveddocumentation
  Task

  ai-dynamo/snapshot#653 · Filter by author ozbarshalom opened on Oct 19, 2026
  

• Snapshots: stage snapshot artifacts from shared storage into a node local-cache for faster restore

  enhancement
  Feature

  ai-dynamo/snapshot#644 · Filter by author ozbarshalom opened on Oct 19, 2026

• Load a conversation KV-cache artifact after restoring an engine checkpoint

  Feature

  ai-dynamo/snapshot#635 · Filter by author dfeigin-nv opened on Oct 18, 2026
  1

• Rootfs tar processing is slow and sensitive to page-cache state

  Feature

  ai-dynamo/snapshot#634 · Filter by author dfeigin-nv opened on Oct 18, 2026

• Retry captures impl

  approved
  Task

  ai-dynamo/snapshot#632 · Filter by author danbar2 opened on Oct 18, 2026 · v0.3.0
  

• [DOC]: Kubernetes walkthrough on DGX Spark (GB10, arm64)

  approveddocumentation
  Task

  ai-dynamo/snapshot#624 · Filter by author alessandro-festa opened on Oct 15, 2026
  2

  

• Restored vLLM pod answers health before the EngineCore's CUDA restore finishes; a request then kills the engine

  documentationgood first issue
  Task

  ai-dynamo/snapshot#623 · Filter by author alessandro-festa opened on Oct 15, 2026
  1

  

• Make custom storage work automatically on kernel-mode drivers older than 615

  Task

  ai-dynamo/snapshot#618 · Filter by author dfeigin-nv opened on Oct 15, 2026
  

• Add AI-slop detection to issues and PRs

  approvedgood first issue
  Task

  ai-dynamo/snapshot#616 · Filter by author dfeigin-nv opened on Oct 15, 2026
  

• Integrate coordinated checkpoint and restore into the node agent

  enhancement
  Task

  ai-dynamo/snapshot#614 · Filter by author julienmancuso opened on Oct 14, 2026
  

• Implement the set checkpoint and restore controllers

  enhancement
  Task

  ai-dynamo/snapshot#613 · Filter by author julienmancuso opened on Oct 14, 2026
  

• Add PodSetSnapshot and PodSetRestore APIs

  enhancement
  Task

  ai-dynamo/snapshot#612 · Filter by author julienmancuso opened on Oct 14, 2026
  

• Support cross-node FABRIC resource reconstruction

  enhancement
  Task

  ai-dynamo/snapshot#611 · Filter by author julienmancuso opened on Oct 14, 2026
  

• Implement the Snapshot coordinator

  enhancement
  Task

  ai-dynamo/snapshot#610 · Filter by author julienmancuso opened on Oct 14, 2026