nsaspy/llm-log
1
0
Fork 0
Code Issues 34 Pull requests 19 Projects Releases Packages Wiki Activity Actions

Unify z.AI and GPT subscription quotas for desktop telemetry #82

Open
nsaspy wants to merge 7 commits from feat/79-quota-telemetry into main
pull from: feat/79-quota-telemetry
merge into: nsaspy:main
Conversation 1 Commits 7 Files changed 10 +799 -2
nsaspy commented 2026-09-13 02:36:34 +00:00
Owner
Copy link

Implements #79's provider-reported quota backend. Qtile consumer: lost-rob0t/dotfiles#237.

Behavior

  • Normalize z.AI and GPT provider-reported percentages USED, window durations, reset times and plan metadata. No hard-coded plan capacities, token-to-credit guesses or fabricated zero/unlimited states.
  • z.AI fixed authenticated quota endpoint supports reported coding token/credit windows, including five-hour and weekly. Unknown shapes remain unknown; MCP time limits are not relabelled as coding quota.
  • GPT uses supported Codex App Server account/read and account/rateLimits/read, retaining rateLimitsByLimitId and using the existing ChatGPT login. Scope is Codex, NOT every ChatGPT chat/voice/model limit. No model thread/turn is started.
  • A bounded asyncio queue actor owns last-good observations. Bounded provider requests/RPCs, retry backoff, safe errors and no credentials in published state.
  • llm-log-quotas exposes cached GET /api/v1/quotas on loopback:8788 and an atomic 0600 XDG cache for desktop widgets. No provider requests on API GET; browser origins/unrecognized Host names rejected.
  • Home Manager service and runtime-only credentials/config paths are independently selectable from the forwarding runtime. Operational docs remain Org-mode.

Executed verification

Final exact head: 2feb4c6bf8.

  • GREEN Quota telemetry run 34733190676: Nix package/full Python tests, CLI entrypoint, Nix syntax and evaluated Home Manager proxy/expert/quota service contract.
  • GREEN Analytics API run 34733190651.
  • Decoded package output: 76 tests, OK with two existing expert-runtime skips. All 20 added quota tests pass, including normalization, plans, stale/reset semantics, actor/cache/API and fake-Codex subprocess handshake.
  • Test-first missing-module RED executed before implementation. CI fixture follow-ups use sys.executable inside Nix and correctly evaluate Home Manager conditional service definitions; checks were not removed.

Boundaries

No live subscription credentials, user desktop access, Home Manager activation, Forgejo synchronization or deployment occurred. Authentication failures are unavailable, never 0%. z.AI unit mapping is source-observed rather than a published stable schema guarantee. This PR does not certify or merge the separate Woo transport/rate stack; attempted repair #81 remains draft after the full transport timeout.

Implements #79's provider-reported quota backend. Qtile consumer: lost-rob0t/dotfiles#237. ## Behavior - Normalize z.AI and GPT provider-reported percentages USED, window durations, reset times and plan metadata. No hard-coded plan capacities, token-to-credit guesses or fabricated zero/unlimited states. - z.AI fixed authenticated quota endpoint supports reported coding token/credit windows, including five-hour and weekly. Unknown shapes remain unknown; MCP time limits are not relabelled as coding quota. - GPT uses supported Codex App Server account/read and account/rateLimits/read, retaining rateLimitsByLimitId and using the existing ChatGPT login. Scope is Codex, NOT every ChatGPT chat/voice/model limit. No model thread/turn is started. - A bounded asyncio queue actor owns last-good observations. Bounded provider requests/RPCs, retry backoff, safe errors and no credentials in published state. - llm-log-quotas exposes cached GET /api/v1/quotas on loopback:8788 and an atomic 0600 XDG cache for desktop widgets. No provider requests on API GET; browser origins/unrecognized Host names rejected. - Home Manager service and runtime-only credentials/config paths are independently selectable from the forwarding runtime. Operational docs remain Org-mode. ## Executed verification Final exact head: 2feb4c6bf8ed8801b7cb864b68c4a9c61565ffd8. - GREEN Quota telemetry run 34733190676: Nix package/full Python tests, CLI entrypoint, Nix syntax and evaluated Home Manager proxy/expert/quota service contract. - GREEN Analytics API run 34733190651. - Decoded package output: 76 tests, OK with two existing expert-runtime skips. All 20 added quota tests pass, including normalization, plans, stale/reset semantics, actor/cache/API and fake-Codex subprocess handshake. - Test-first missing-module RED executed before implementation. CI fixture follow-ups use sys.executable inside Nix and correctly evaluate Home Manager conditional service definitions; checks were not removed. ## Boundaries No live subscription credentials, user desktop access, Home Manager activation, Forgejo synchronization or deployment occurred. Authentication failures are unavailable, never 0%. z.AI unit mapping is source-observed rather than a published stable schema guarantee. This PR does not certify or merge the separate Woo transport/rate stack; attempted repair #81 remains draft after the full transport timeout.
ci: archive consumer sources without checking out unrelated gitlinks
Some checks failed
Quota telemetry / contracts (push) Failing after 42s
Details
e33593feb0
feat: unify provider-reported z.AI and GPT subscription quota telemetry (#79)
Some checks failed
Quota telemetry / contracts (push) Failing after 4s
Details
30c9d5f0b6
test: evaluate conditional service definitions through Home Manager-shaped submodules
Some checks failed
Quota telemetry / contracts (push) Failing after 3s
Details
Quota telemetry / contracts (pull_request) Failing after 3s
Details
Analytics API / test (pull_request) Failing after 4s
Details
Analytics API / package (pull_request) Has been skipped
Details
2feb4c6bf8
nsaspy commented 2026-09-21 20:13:17 +00:00
Author
Owner
Copy link

Final quota backend head: 2feb4c6bf8.

Verified GREEN at this exact head:

  • Quota telemetry run 34733190676: package/full Python contracts, CLI entrypoint, Nix syntax and evaluated Home Manager proxy/expert/quota service contract all passed.
  • Analytics API run 34733190651: passed.
  • The decoded package test output on this runtime reported 76 tests, OK with two existing expert-runtime skips; the 20 new quota normalization/actor/cache/API/Codex-handshake tests pass.

Two CI fixture issues were repaired without weakening production checks: use sys.executable rather than /usr/bin/env in the subprocess fixture, and model conditional Home Manager services as submodules with context-free comparison needles. Production provider telemetry behavior did not change in those follow-ups.

Dotfiles PR lost-rob0t/dotfiles#237 is pinned to this exact head. No live subscription credential verification, user-host activation, Forgejo synchronization or deployment occurred. GPT remains explicitly Codex quota scope under ChatGPT authentication, not all ChatGPT chat/voice/model usage. The separate gateway transport repair #81 is still blocked and unmerged.

Final quota backend head: 2feb4c6bf8ed8801b7cb864b68c4a9c61565ffd8. Verified GREEN at this exact head: - Quota telemetry run 34733190676: package/full Python contracts, CLI entrypoint, Nix syntax and evaluated Home Manager proxy/expert/quota service contract all passed. - Analytics API run 34733190651: passed. - The decoded package test output on this runtime reported 76 tests, OK with two existing expert-runtime skips; the 20 new quota normalization/actor/cache/API/Codex-handshake tests pass. Two CI fixture issues were repaired without weakening production checks: use sys.executable rather than /usr/bin/env in the subprocess fixture, and model conditional Home Manager services as submodules with context-free comparison needles. Production provider telemetry behavior did not change in those follow-ups. Dotfiles PR lost-rob0t/dotfiles#237 is pinned to this exact head. No live subscription credential verification, user-host activation, Forgejo synchronization or deployment occurred. GPT remains explicitly Codex quota scope under ChatGPT authentication, not all ChatGPT chat/voice/model usage. The separate gateway transport repair #81 is still blocked and unmerged.
Some checks failed
Quota telemetry / contracts (push) Failing after 3s
Details
Quota telemetry / contracts (pull_request) Failing after 3s
Details
Analytics API / test (pull_request) Failing after 4s
Details
Analytics API / package (pull_request) Has been skipped
Details
This pull request has changes conflicting with the target branch.
  • .github/workflows/quota-telemetry.yml
  • llm_log/quotas.py
  • roam/index.org
  • roam/solutions/provider-quota-telemetry.org
  • tests/test_quotas.py
View command line instructions

Manual merge helper

Use this merge commit message when completing the merge manually.

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u origin feat/79-quota-telemetry:feat/79-quota-telemetry
git switch feat/79-quota-telemetry

Merge

Merge the changes and update on Forgejo.

Warning: The "Autodetect manual merge" setting is not enabled for this repository, you will have to mark this pull request as manually merged afterwards.

git switch main
git merge --no-ff feat/79-quota-telemetry
git switch feat/79-quota-telemetry
git rebase main
git switch main
git merge --ff-only feat/79-quota-telemetry
git switch feat/79-quota-telemetry
git rebase main
git switch main
git merge --no-ff feat/79-quota-telemetry
git switch main
git merge --squash feat/79-quota-telemetry
git switch main
git merge --ff-only feat/79-quota-telemetry
git switch main
git merge feat/79-quota-telemetry
git push origin main
Sign in to join this conversation.
Reviewers
No reviewers
Labels
Clear labels   
accessibility
Barrier affecting people with disabilities

  
bug
Something isn't working

  
documentation
Improvements or additions to documentation

  
duplicate
This issue or pull request already exists

  
enhancement
New feature or request

  
good first issue
Good for newcomers

  
help wanted
Extra attention is needed

  
invalid
This doesn't seem right

  
question
Further information is requested

  
wontfix
This will not be worked on

No labels
accessibility
bug
documentation
duplicate
enhancement
good first issue
help wanted
invalid
question
wontfix
Milestone
Clear milestone
No items
No milestone
Projects
Clear projects
No items
No project
Assignees
Clear assignees
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set.

Reference
nsaspy/llm-log!82
No description provided.