Skip to content

Latest commit

 

History

History
129 lines (82 loc) · 3.69 KB

File metadata and controls

129 lines (82 loc) · 3.69 KB

Troubleshooting guide

Problems with the toolkit itself and how to fix them. (For troubleshooting your infrastructure, that's what the toolkit is for — see Use cases.)

Installation

devops-ai: command not found

The console script isn't on your PATH. Confirm the install and your environment:

pip show devops-ai-toolkit
python -m pip install --user devops-ai-toolkit   # if installing per-user

If you installed into a virtualenv, make sure it's activated. You can always invoke via the module entry indirectly through the SDK, or reinstall in the active environment.

Install the API extra: pip install 'devops-ai-toolkit[api]'

devops-ai serve needs FastAPI + Uvicorn, which ship as the optional api extra:

pip install 'devops-ai-toolkit[api]'

requires-python / Python version error

The toolkit needs Python 3.12+. Check with python --version and use a 3.12 interpreter.

Analysis

"No known … error signature matched"

The input was read but didn't match a catalogued pattern. Options:

  • Add a --tech hint so detection isn't the issue: devops-ai analyze app.log --tech kubernetes.
  • Provide more of the raw log rather than a paraphrase.
  • The pattern may simply not be in the knowledge base yet — consider adding a signature, or browse https://devopsaitoolkit.com/blog.

"No input. Pass a file path or pipe content via stdin."

analyze/validate got nothing. Either pass a file path or pipe content and use -:

devops-ai analyze app.log
cat app.log | devops-ai analyze -

File not found

The path argument doesn't exist. Check the path; exit code is 2 for input errors.

Output looks truncated

Inputs are truncated to DEVOPS_AI_MAX_CHARS (default 200,000). Raise it if you must analyze larger text:

export DEVOPS_AI_MAX_CHARS=500000

See Configuration.

Exit codes

Code Meaning
0 Matched / valid
1 No match / validation failed
2 Input error (missing file, empty stdin)

If a script "fails" on a clean run, remember analyze exits 1 when nothing matched — that's a valid outcome, not an error.

AI enrichment

Enrichment was skipped

You passed --enrich / enrich=True but no provider is available, so you got a low-severity warning and deterministic-only results. Set a provider and its key:

export DEVOPS_AI_PROVIDER=anthropic
export ANTHROPIC_API_KEY=sk-ant-...

Check readiness via the API:

curl -s localhost:8000/health | jq '{provider, provider_available}'

See AI providers.

Provider/network errors

If a provider call fails, the engine logs a warning and returns the deterministic result — analysis is never broken by an LLM failure. Verify the key, model name, network/proxy, and DEVOPS_AI_TIMEOUT.

Wrong model is being used

Override per provider, e.g. DEVOPS_AI_ANTHROPIC_MODEL, DEVOPS_AI_OPENAI_MODEL, DEVOPS_AI_GEMINI_MODEL, DEVOPS_AI_OLLAMA_MODEL. See Configuration.

REST API

Can't reach the server

Defaults to 127.0.0.1:8000. To accept remote connections bind to 0.0.0.0:

devops-ai serve --host 0.0.0.0 --port 8000

Remember the API ships no auth — front it with your own. See Security.

Where are the docs?

Swagger UI at /docs, ReDoc at /redoc, OpenAPI JSON at /openapi.json.

Still stuck?