github-starred/ollama
2
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2026-05-07 00:22:43 -05:00
Code Issues 14.5k Packages Projects Releases 120 Wiki Activity

[GH-ISSUE #14369] Suggestion: RAG / LLM pipeline failure-mode checklist (WFGY ProblemMap) for Ollama users #35097

New Issue
Open
opened 2026-04-22 19:18:46 -05:00 by GiteaMirror · 3 comments
GiteaMirror commented 2026-04-22 19:18:46 -05:00
Owner
Copy Link

Originally created by @onestardao on GitHub (Feb 23, 2026).
Original GitHub issue: https://github.com/ollama/ollama/issues/14369

Hi, and thanks for Ollama – it has become a key part of many local-first LLM and RAG stacks.

I maintain an MIT-licensed open-source project called WFGY (~1.5k GitHub stars).
One of its core components is a 16-problem “ProblemMap” for RAG / LLM pipelines, which catalogues common failure modes across:

  • data ingestion and chunking
  • embeddings and vector stores
  • retrievers, rerankers, and routing
  • LLM prompting, tools, and orchestration
  • evaluation gaps and guardrails

ProblemMap overview (English docs, with code examples and checklists):
https://github.com/onestardao/WFGY/blob/main/ProblemMap/README.md

This checklist is already referenced or integrated in several research-oriented projects and curated lists, including for example:

  • ToolUniverse from the Harvard MIMS Lab (RAG robustness / debugging entry)
  • A multimodal RAG survey from QCRI LLM Lab (Qatar Computing Research Institute)
  • Rankify from the University of Innsbruck (RAG evaluation and debugging)
  • community “awesome data science / ML” lists that focus on production RAG

In practice, many people are now building RAG pipelines where Ollama is the model backend, while the rest of the stack is LangChain / LlamaIndex / Haystack / custom code etc. Most of the failures they hit are not model bugs, but exactly the ingestion / embeddings / retrieval / guardrail problems described in the 16-problem checklist above.

Suggestion

Would you be open to adding this as an external resource for Ollama users who are building RAG pipelines?

For example, one of the following options:

  • a short link in the docs under a “RAG troubleshooting / failure modes” or “best practices” section, or
  • a one-line entry in the README near the RAG & Knowledge Bases / Observability sections as a “RAG / LLM pipeline failure-mode checklist”.

If this sounds useful, I am happy to open a very small PR that only adds a single bullet-point link, following your existing formatting. If you prefer not to add external checklists in the main README, totally understood – in that case I’ll just keep pointing Ollama users to this resource from my own side.

Thanks again for making Ollama such a solid base for local-first RAG and agent systems.

Originally created by @onestardao on GitHub (Feb 23, 2026). Original GitHub issue: https://github.com/ollama/ollama/issues/14369 Hi, and thanks for Ollama – it has become a key part of many local-first LLM and RAG stacks. I maintain an MIT-licensed open-source project called **WFGY** (~1.5k GitHub stars). One of its core components is a **16-problem “ProblemMap” for RAG / LLM pipelines**, which catalogues common failure modes across: - data ingestion and chunking - embeddings and vector stores - retrievers, rerankers, and routing - LLM prompting, tools, and orchestration - evaluation gaps and guardrails ProblemMap overview (English docs, with code examples and checklists): https://github.com/onestardao/WFGY/blob/main/ProblemMap/README.md This checklist is already referenced or integrated in several research-oriented projects and curated lists, including for example: - ToolUniverse from the Harvard MIMS Lab (RAG robustness / debugging entry) - A multimodal RAG survey from QCRI LLM Lab (Qatar Computing Research Institute) - Rankify from the University of Innsbruck (RAG evaluation and debugging) - community “awesome data science / ML” lists that focus on production RAG In practice, many people are now building RAG pipelines where **Ollama is the model backend**, while the rest of the stack is LangChain / LlamaIndex / Haystack / custom code etc. Most of the failures they hit are not model bugs, but exactly the ingestion / embeddings / retrieval / guardrail problems described in the 16-problem checklist above. **Suggestion** Would you be open to adding this as an external resource for Ollama users who are building RAG pipelines? For example, one of the following options: - a short link in the docs under a “RAG troubleshooting / failure modes” or “best practices” section, or - a one-line entry in the README near the RAG & Knowledge Bases / Observability sections as a “RAG / LLM pipeline failure-mode checklist”. If this sounds useful, I am happy to open a very small PR that only adds a single bullet-point link, following your existing formatting. If you prefer not to add external checklists in the main README, totally understood – in that case I’ll just keep pointing Ollama users to this resource from my own side. Thanks again for making Ollama such a solid base for local-first RAG and agent systems.
GiteaMirror commented 2026-04-22 19:18:49 -05:00
Author
Owner
Copy Link

@lingfan36 commented on GitHub (Feb 25, 2026):

👋 你好！

这是 API 相关的问题。请参考我们的 API 文档：

🔗 API 参考文档: https://ollamahub.space/pages/docs/api/

文档中包含完整的 REST API 说明和示例代码。

由 OllamaHub 自动生成

<!-- gh-comment-id:3957147780 --> @lingfan36 commented on GitHub (Feb 25, 2026): 👋 你好！ 这是 API 相关的问题。请参考我们的 API 文档： 🔗 **API 参考文档**: https://ollamahub.space/pages/docs/api/ 文档中包含完整的 REST API 说明和示例代码。 --- *由 OllamaHub 自动生成*
GiteaMirror commented 2026-04-22 19:18:50 -05:00
Author
Owner
Copy Link

@lingfan36 commented on GitHub (Feb 25, 2026):

👋 你好！

这是 API 相关的问题。请参考我们的 API 文档：

🔗 API 参考文档: https://ollamahub.space/pages/docs/api/

文档中包含完整的 REST API 说明和示例代码。

由 OllamaHub 自动生成

<!-- gh-comment-id:3957147838 --> @lingfan36 commented on GitHub (Feb 25, 2026): 👋 你好！ 这是 API 相关的问题。请参考我们的 API 文档： 🔗 **API 参考文档**: https://ollamahub.space/pages/docs/api/ 文档中包含完整的 REST API 说明和示例代码。 --- *由 OllamaHub 自动生成*
GiteaMirror commented 2026-04-22 19:18:51 -05:00
Author
Owner
Copy Link

@onestardao commented on GitHub (Mar 3, 2026):

Hi, just a gentle follow-up in case this got buried.

I completely understand if this isn’t a priority right now. If adding an external checklist isn’t aligned with the repo direction, no worries at all.

If it’s helpful for Ollama users building RAG stacks, I’m still happy to open a very small PR that follows your existing docs format.

Thanks again for your time.

<!-- gh-comment-id:3989675460 --> @onestardao commented on GitHub (Mar 3, 2026): Hi, just a gentle follow-up in case this got buried. I completely understand if this isn’t a priority right now. If adding an external checklist isn’t aligned with the repo direction, no worries at all. If it’s helpful for Ollama users building RAG stacks, I’m still happy to open a very small PR that follows your existing docs format. Thanks again for your time.
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
amd
api
app
bug
build
cli
cloud
compatibility
context-length
create
docker
documentation
embeddings
feature request
feedback wanted
good first issue
gpt-oss
gpu
harmony
help wanted
image
install
intel
js
launch
linux
macos
memory
mlx
model
needs more info
networking
nvidia
ollama.com
performance
pull-request
Mirrored from GitHub Pull Request
 python
question
registry
rendering
thinking
tools
top
vulkan
windows
wsl
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
GiteaMirror ninjasurge
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: github-starred/ollama#35097
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?