Files
Brisbanehuang 20ad5ec506 feat(scheduler): per-group profit control for token account admission
Group pricing (rate multiplier, peak windows, per-user overrides) and
account cost (accounts.rate_multiplier) already live side by side, but
nothing stops the scheduler from handing a request to an account whose
cost multiplier exceeds what the group's pricing can profitably serve.
Add an opt-in per-group profit gate that filters scheduling candidates
by a margin rule, while ordering, scoring, stickiness and breakers keep
working unchanged among qualified accounts.

Admission rule: an account qualifies iff U <= D * (1 - min_margin -
safety_buffer) within a small relative epsilon, where U is
accounts.rate_multiplier (0 is legal; missing/negative/NaN/Inf are
conservatively rejected as invalid) and D is the requester's effective
downstream multiplier (user-group override ?? group default, times the
group peak factor) frozen at the request's pricing instant.

- groups gain profit_control_enabled / profit_min_margin /
  profit_safety_buffer (migration 191); the durable auth-cache
  invalidation trigger additionally watches the profit and pricing
  columns (migration 192) so out-of-band group edits cannot leave
  stale auth snapshots; GetByKeyForAuth explicitly projects the new
  columns and the API-key auth snapshot version is bumped to force a
  refresh of pre-existing snapshots
- request-level pricing instant: token entry points install pricingAt
  into ctx; the profit threshold D and the RecordUsage peak factor
  read the same instant, so one request never changes price mid-flight
  across waits/retries/failover (media and unwired paths keep the
  existing record-time semantics)
- the gate covers token requests on openai, anthropic, gemini, grok
  and antigravity groups: OpenAI-family handlers via
  WithOpenAIRequestPricingContext (responses incl. WS bridge, chat
  completions, messages, embeddings, alpha search), the shared gateway
  via WithGatewayTokenRequestPricing (messages, chat completions,
  responses, gemini model actions); composite groups cannot enable it
  directly; image/video/models/usage/count_tokens stay ungated and an
  explicit image-generation intent suppresses the gate end to end
- post-slot recheck: after a slot is acquired the account is re-read
  via SchedulerSnapshotService.GetAccount (scheduler cache, then DB;
  only when both fail the check fails open with WARN + metric); a
  vetoed account releases its slot and joins the request's exclusion
  set for reselection; sticky bindings are written only after the
  final check passes, and an over-threshold sticky account is skipped,
  not deleted, so it comes back once its rate recovers
- sticky-session cache contract: GatewayCache.GetSessionAccountID now
  returns ErrStickySessionNotFound on a miss (mapped from redis.Nil in
  the repository implementation, mirroring ErrRefreshTokenNotFound) so
  the profit sticky path can distinguish "no binding yet" from a real
  read failure without importing the cache driver in service code
- cross-group re-entry (composite parent -> member group) resolves the
  gate against the member group and clears a stale parent gate instead
  of letting a foreign threshold veto accounts
- per-platform/group activity counters (installs, threshold vetoes,
  invalid-rate vetoes, refresh failures) for observability
- admin UI: profit-control section on the five platforms' group forms
  with percent input, validation and platform-switch reset; group
  create/update/duplicate normalize and validate the config at a
  single choke point
- cmd/profit-preview: offline what-if tool that replays the production
  admission semantics over an exported config/account/override/model
  dump, reports per-model admitted-account counts under the default
  and the worst-case (lowest user override) D, and surfaces probe-sync
  staleness as warnings without affecting admission

Tests: service unit coverage for gate resolution/veto/threshold
epsilon/pricing instant/suppress marker/scheduler filtering and
post-slot recheck (incl. -race on the profit surface), unit-tagged
handler slot-recheck and capability-mapping regressions, sqlmock and
real-PostgreSQL integration regressions for the GetByKeyForAuth
projection and the migration-192 trigger watch list, API contract
update, and frontend specs for the five-platform form helpers.
2026-08-01 22:39:31 +08:00

147 lines
5.2 KiB
Go

package service
import (
"testing"
"time"
"github.com/stretchr/testify/require"
)
func TestPreviewProfitAdmissionUsesAccountRatesAndPreinitializesModels(t *testing.T) {
now := time.Now()
group := profitControlTestGroup(50, 0.2, 0)
group.Name = "VIP-preview"
cheap := profitControlTestAccountWithRate(
upstreamCostTestAccount(1, UpstreamBillingProbeStatusOK, 1.0, now.Add(-3*time.Hour), 30*time.Minute),
0.5,
)
cheap.Name = "cheap"
cheap.Extra[UpstreamBillingRateSyncEnabledExtraKey] = true
cheap.Credentials = map[string]any{"model_mapping": map[string]any{"gpt-sol": "gpt-sol"}}
boundary := profitControlTestAccountWithRate(
upstreamCostTestAccount(2, UpstreamBillingProbeStatusOK, 0.8, now.Add(-time.Minute), 30*time.Minute),
0.8,
)
boundary.Name = "boundary"
boundary.Credentials = map[string]any{"model_mapping": map[string]any{
"gpt-sol": "gpt-sol",
"gpt-luna": "gpt-luna",
}}
expensive := profitControlTestAccountWithRate(
upstreamCostTestAccount(3, UpstreamBillingProbeStatusOK, 0.2, now.Add(-time.Minute), 30*time.Minute),
1.0,
)
expensive.Name = "expensive"
invalid := upstreamCostTestOAuthAccount(4)
invalid.Name = "invalid"
reports := PreviewProfitAdmission([]ProfitPreviewGroupInput{{
Group: group,
Accounts: []*Account{cheap, boundary, expensive, invalid},
UserOverrides: map[int64]float64{40: 0.5},
Models: []string{"gpt-sol", "gpt-luna", "gpt-no-account"},
}}, now)
require.Len(t, reports, 1)
report := reports[0]
require.True(t, report.EffectiveGate)
require.InDelta(t, 1.0, report.DefaultD, 1e-12)
require.InDelta(t, 0.8, report.ThresholdDefault, 1e-12)
require.InDelta(t, 0.5, report.MinEffectiveD, 1e-12)
require.InDelta(t, 0.4, report.ThresholdMinD, 1e-12)
byID := map[int64]ProfitPreviewAccountVerdict{}
for _, verdict := range report.Verdicts {
byID[verdict.AccountID] = verdict
}
require.Equal(t, ProfitPreviewClassAdmitted, byID[cheap.ID].Class)
require.Equal(t, ProfitPreviewRateSourceUpstreamProbe, byID[cheap.ID].RateSource)
require.Contains(t, byID[cheap.ID].Warnings, ProfitPreviewWarningProbeStale)
require.True(t, byID[cheap.ID].RejectedUnderMinD)
require.Equal(t, ProfitPreviewClassAdmitted, byID[boundary.ID].Class, "U == 阈值按 epsilon 语义准入")
require.Equal(t, ProfitPreviewClassRejectedThreshold, byID[expensive.ID].Class)
require.Contains(t, byID[expensive.ID].Warnings, ProfitPreviewWarningManualRateOne)
require.Equal(t, ProfitPreviewClassRejectedInvalidRate, byID[invalid.ID].Class)
require.Equal(t, 2, report.RemainingByModel["gpt-sol"])
require.Equal(t, 1, report.RemainingByModel["gpt-luna"])
require.Equal(t, 0, report.RemainingByModel["gpt-no-account"])
_, present := report.RemainingByModel["gpt-no-account"]
require.True(t, present, "全部账号均不支持时也必须显式保留 0,供 CLI 发出警告")
}
func TestPreviewProfitAdmissionAssumeEnabled(t *testing.T) {
now := time.Now()
group := profitControlTestGroup(51, 0, 0)
group.Platform = PlatformOpenAI
group.RateMultiplier = 0.5
group.ProfitControlEnabled = false
cheapAccount := profitControlTestAccountWithRate(
upstreamCostTestAccount(1, UpstreamBillingProbeStatusOK, 0.9, now.Add(-time.Minute), 30*time.Minute),
0.2,
)
expensiveAccount := profitControlTestAccountWithRate(
upstreamCostTestAccount(2, UpstreamBillingProbeStatusOK, 0.2, now.Add(-time.Minute), 30*time.Minute),
0.9,
)
withoutAssume := PreviewProfitAdmission([]ProfitPreviewGroupInput{{
Group: group,
Accounts: []*Account{cheapAccount, expensiveAccount},
Models: []string{"gpt-test"},
}}, now)[0]
require.False(t, withoutAssume.EffectiveGate)
require.Equal(t, ProfitPreviewClassAdmitted, withoutAssume.Verdicts[0].Class)
require.Equal(t, ProfitPreviewClassAdmitted, withoutAssume.Verdicts[1].Class)
withAssume := PreviewProfitAdmission([]ProfitPreviewGroupInput{{
Group: group,
Accounts: []*Account{cheapAccount, expensiveAccount},
Models: []string{"gpt-test"},
AssumeEnabled: true,
}}, now)[0]
require.True(t, withAssume.EffectiveGate)
require.True(t, withAssume.AssumedEnabled)
byID := map[int64]ProfitPreviewAccountVerdict{}
for _, verdict := range withAssume.Verdicts {
byID[verdict.AccountID] = verdict
}
require.Equal(t, ProfitPreviewClassAdmitted, byID[cheapAccount.ID].Class,
"账号倍率 0.2 <= 阈值 0.5,探测快照的高倍率不参与准入判断")
require.Equal(t, ProfitPreviewClassRejectedThreshold, byID[expensiveAccount.ID].Class,
"账号倍率 0.9 > 阈值 0.5,探测快照的低倍率不能替代账号倍率")
}
func TestPreviewProfitAdmissionSupportsFivePlatforms(t *testing.T) {
for i, platform := range []string{
PlatformOpenAI,
PlatformAnthropic,
PlatformGemini,
PlatformGrok,
PlatformAntigravity,
} {
group := profitControlTestGroup(int64(100+i), 0, 0)
group.Platform = platform
rate := 0.2
account := &Account{
ID: int64(200 + i),
Platform: platform,
RateMultiplier: &rate,
}
report := PreviewProfitAdmission([]ProfitPreviewGroupInput{{
Group: group,
Accounts: []*Account{account},
Models: []string{"model"},
}}, time.Now())[0]
require.True(t, report.EffectiveGate, platform)
require.Equal(t, ProfitPreviewClassAdmitted, report.Verdicts[0].Class, platform)
}
}