feat: runnable onboarding quickstart, recipes, and fixtures (issue #32)#47
Open
levleontiev wants to merge 65 commits intomainfrom
Open
feat: runnable onboarding quickstart, recipes, and fixtures (issue #32)#47levleontiev wants to merge 65 commits intomainfrom
levleontiev wants to merge 65 commits intomainfrom
Conversation
…hmark link (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
…ig shapes (issue #32)
Failover between providers is a client-side pattern, not a Fairvisor feature. Replace with a focused circuit-breaker recipe that demonstrates the actual cost-spike auto-shutdown capability: - spend_rate_threshold_per_minute triggers full traffic block - auto_reset_after_minutes provides hands-free cooldown - per-org TPM limit continues to run in parallel
… + ngx.shared.dict refs
…x, ip:address, curl formatting, Architecture ToC
…loop_detection config
…ngside existing gateways
…-free hash (85k ASNs)
… support - Fix _simple_word_estimate to properly skip false-positive "content" key matches (e.g. when key appears inside a string value) by verifying the value separator (: "...") before counting characters - Add _extract_max_tokens() to parse max_tokens / max_completion_tokens from raw request body when request_context.max_tokens is not set, enabling accurate budget reservation for OpenAI-compatible payloads Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…vars - Remove trailing whitespace on blank lines (lines 201, 205) - Replace unused s, e variables with _ in _extract_max_tokens() Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add two scenarios to exercise the uncovered branches in _extract_max_tokens: - body contains "max_tokens" field → uses it (covers return tonumber(val) branch) - body has no max_tokens/max_completion_tokens → falls back to default_max_completion (covers return nil branch) Fixes coverage regression (91.61% < 91.63% threshold). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Implements #32 — packages the edge repo as a runnable onboarding funnel.
What's added:
examples/quickstart/—docker compose up -dquickstart (wrapper mode, standalone). Reusable as the base fore2e-smokeCI. Demonstrates/openai/v1/chat/completionswithBearer CLIENT_JWT:UPSTREAM_KEYcomposite auth.examples/recipes/team-budgets/— per-team JWT TPM + 30-day cost budget with staged warn/throttle/rejectexamples/recipes/runaway-agent-guard/— loop detector + TPM guard keyed onjwt:agent_idexamples/recipes/provider-failover/— dual-provider policies (OpenAI + circuit breaker, Anthropic fallback)fixtures/— 10 canonical artifacts:normal_request.json,over_limit_request.json,anthropic_normal_request.jsonallow_response.json(proves auth-header stripping: noAuthorization/x-api-key/x-goog-api-keyin response)reject_tpm_exceeded.json,reject_tpd_exceeded.json,reject_prompt_too_large.jsonreject_openai.json,reject_anthropic.json,reject_gemini.json(provider-native bodies)fairvisor/benchmarkSpec 019 / wrapper mode coverage:
FAIRVISOR_MODE=wrapperand provider-prefixed paths (/openai/v1/...)allow_response.jsonexplicitly documents auth-header strippingreject_anthropic.jsonandreject_gemini.jsonprove provider-native error bodiesAcceptance criteria met:
docker compose up -d+ curl to/openai/v1/chat/completionswith over-limit fixture → 429 matchingfixtures/reject_tpm_exceeded.json✓examples/read as deployable recipes, not disconnected policy fragments ✓fairvisor/benchmarkinstead of duplicating methodology ✓Downstream: Fixture paths and compose structure are stable — ready to coordinate with
fairvisor/documentation#8.Test plan
docker compose -f examples/quickstart/docker-compose.yml up -dstarts healthycurlto/openai/v1/chat/completionswithover_limit_request.jsonreturns 429 with body matchingfixtures/reject_tpm_exceeded.jsonnormal_request.jsonreturns 200Authorization,x-api-key,x-goog-api-keyfairvisor validate)🤖 Generated with Claude Code