Skip to content

Add spend-caps skill; document LLM usage analytics - #29

Open
ojaragodwin wants to merge 1 commit into
mainfrom
add-spend-caps-skill
Open

Add spend-caps skill; document LLM usage analytics#29
ojaragodwin wants to merge 1 commit into
mainfrom
add-spend-caps-skill

Conversation

@ojaragodwin

Copy link
Copy Markdown

Summary

  • New iblai-api-spend-caps skill documenting the six LLM spend-cap endpoints: tenant / agent / user-per-agent cap CRUD (PUT upsert, DELETE) plus the learner-safe status read. Covers the three scopes, interval/enforcement/threshold semantics, read-only ClickHouse-reconciled counters, and the 429-on-block behavior (vs credits' 402).
  • Extend iblai-api-analytics with the fine-grained LLM usage endpoint ({dm_url}/api/analytics/llm-usage/): metrics / observations / traces resources, whitelisted measure/aggregation/dimension vocab, get-by-id, and the {resource, data, meta} envelope.

Both are pinned to the backend source (spend_cap_views, SpendCapSerializer, the SpendCap model, utils/spend_caps, and the langfuse analytics view) and follow the existing SKILL.md format.

Test plan

  • Review the two SKILL.md files for endpoint/path/enum accuracy against the ibl-dm-pro backend.

Add iblai-api-spend-caps documenting the six LLM spend-cap endpoints
(tenant/agent/user-per-agent caps + learner-safe status), and extend
iblai-api-analytics with the fine-grained LLM usage endpoint
(per-model/-agent/-user cost, tokens, latency).
Sign up for freeto join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant

@ojaragodwin