feat(acceptance): 建立同构验收与弹性容量体系

实现本地三节点 K3s 同构环境、脱敏生产快照、Gemini 图片和多参考图视频协议模拟、统一验收报告及故障注入。\n\n新增 Worker 容量控制器、资源与连接预算、任务恢复保护,并将生产验收拆分为 validation 执行和人工 CAS 放量。\n\n验证包括 Go 全量测试、PostgreSQL HTTP 集成测试、go vet、OpenAPI、ShellCheck、前端检查、迁移及发布脚本测试。
This commit is contained in:
2026-07-31 18:02:24 +08:00
parent 93d36d0e55
commit e05922b0f4
94 changed files with 12379 additions and 826 deletions
+500
View File
@@ -0,0 +1,500 @@
package main
import (
"context"
"crypto/rand"
"encoding/hex"
"encoding/json"
"errors"
"flag"
"fmt"
"os"
"path/filepath"
"regexp"
"strings"
"time"
"github.com/easyai/easyai-ai-gateway/apps/api/internal/acceptancesnapshot"
"github.com/easyai/easyai-ai-gateway/apps/api/internal/auth"
"github.com/easyai/easyai-ai-gateway/apps/api/internal/store"
)
var (
fullSHAPattern = regexp.MustCompile(`^[0-9a-f]{40}$`)
digestPattern = regexp.MustCompile(`^sha256:[0-9a-f]{64}$`)
)
type options struct {
clusterID string
releaseSHA string
apiImageDigest string
workerDigest string
emulatorBaseURL string
callbackURL string
output string
identityShards int
}
type runtimeFile struct {
SchemaVersion string `json:"schemaVersion"`
LocalClusterID string `json:"localClusterId"`
RunID string `json:"runId"`
RunToken string `json:"runToken"`
APIKeys []string `json:"apiKeys"`
Participants []runtimeParticipant `json:"participants"`
GeminiModel string `json:"geminiModel"`
VideoModel string `json:"videoModel"`
EmulatorBaseURL string `json:"emulatorBaseUrl"`
CallbackURL string `json:"callbackUrl"`
ReleaseSHA string `json:"releaseSha"`
APIImageDigest string `json:"apiImageDigest"`
WorkerImageDigest string `json:"workerImageDigest"`
SnapshotConfigHash string `json:"snapshotConfigHash"`
SnapshotSHA256 string `json:"snapshotSha256"`
CreatedAt time.Time `json:"createdAt"`
}
type runtimeParticipant struct {
APIKeyID string `json:"apiKeyId"`
UserID string `json:"userId"`
}
func main() {
opts, err := parseOptions()
if err != nil {
fmt.Fprintln(os.Stderr, "acceptance bootstrap:", err)
os.Exit(64)
}
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Minute)
defer cancel()
if err := run(ctx, opts); err != nil {
fmt.Fprintln(os.Stderr, "acceptance bootstrap:", err)
os.Exit(1)
}
}
func parseOptions() (options, error) {
var opts options
flag.StringVar(&opts.clusterID, "local-cluster-id", "", "required local cluster marker")
flag.StringVar(&opts.releaseSHA, "release-sha", "", "full source Git SHA")
flag.StringVar(&opts.apiImageDigest, "api-image-digest", "", "immutable API image digest")
flag.StringVar(&opts.workerDigest, "worker-image-digest", "", "immutable Worker image digest")
flag.StringVar(&opts.emulatorBaseURL, "emulator-base-url", "", "in-cluster protocol emulator URL")
flag.StringVar(&opts.callbackURL, "callback-url", "", "in-cluster callback collector URL")
flag.StringVar(&opts.output, "output", "", "private runtime output file")
flag.IntVar(&opts.identityShards, "identity-shards", 32, "isolated acceptance identities")
flag.Parse()
opts.clusterID = strings.TrimSpace(opts.clusterID)
opts.releaseSHA = strings.ToLower(strings.TrimSpace(opts.releaseSHA))
opts.apiImageDigest = strings.ToLower(strings.TrimSpace(opts.apiImageDigest))
opts.workerDigest = strings.ToLower(strings.TrimSpace(opts.workerDigest))
opts.emulatorBaseURL = strings.TrimRight(strings.TrimSpace(opts.emulatorBaseURL), "/")
opts.callbackURL = strings.TrimSpace(opts.callbackURL)
if opts.clusterID == "" || !fullSHAPattern.MatchString(opts.releaseSHA) ||
!digestPattern.MatchString(opts.apiImageDigest) || !digestPattern.MatchString(opts.workerDigest) {
return options{}, errors.New("local cluster ID, full release SHA, and immutable image digests are required")
}
if opts.emulatorBaseURL == "" || opts.callbackURL == "" {
return options{}, errors.New("emulator and callback URLs are required")
}
if opts.identityShards < 1 || opts.identityShards > 128 {
return options{}, errors.New("identity shards must be between 1 and 128")
}
if strings.TrimSpace(opts.output) == "" {
return options{}, errors.New("private runtime output path is required")
}
if flag.NArg() != 0 {
return options{}, errors.New("unexpected positional arguments")
}
return opts, nil
}
func run(ctx context.Context, opts options) error {
databaseURL := strings.TrimSpace(os.Getenv("AI_GATEWAY_DATABASE_URL"))
if databaseURL == "" {
return errors.New("AI_GATEWAY_DATABASE_URL is required")
}
database, err := store.ConnectWithMaxConns(ctx, databaseURL, 4)
if err != nil {
return err
}
defer database.Close()
if err := verifyLocalDatabase(ctx, database, opts.clusterID); err != nil {
return err
}
snapshotConfigHash, snapshotSHA, err := importedSnapshot(ctx, database)
if err != nil {
return err
}
groupID, tenantID, err := ensureAcceptanceIdentityDomain(ctx, database)
if err != nil {
return err
}
participants := make([]runtimeParticipant, 0, opts.identityShards)
apiKeys := make([]string, 0, opts.identityShards)
for ordinal := 0; ordinal < opts.identityShards; ordinal++ {
userID, err := ensureAcceptanceUser(ctx, database, groupID, tenantID, ordinal)
if err != nil {
return err
}
keyID, secret, err := ensureAcceptanceAPIKey(ctx, database, userID, groupID, tenantID, ordinal)
if err != nil {
return err
}
if _, err := database.SetUserWalletBalance(ctx, store.WalletBalanceAdjustmentInput{
GatewayUserID: userID,
Currency: "resource",
BalanceText: "1000000000",
Reason: "local acceptance isolated wallet",
}); err != nil && !errors.Is(err, store.ErrWalletBalanceUnchanged) {
return err
}
participants = append(participants, runtimeParticipant{APIKeyID: keyID, UserID: userID})
apiKeys = append(apiKeys, secret)
}
if err := ensureAcceptanceAccessRules(ctx, database, groupID); err != nil {
return err
}
geminiModel, videoModel, err := selectedModels(ctx, database)
if err != nil {
return err
}
runToken, err := randomToken()
if err != nil {
return err
}
run, err := database.CreateAcceptanceRun(ctx, store.CreateAcceptanceRunInput{
ReleaseSHA: opts.releaseSHA,
APIImageDigest: opts.apiImageDigest,
WorkerImageDigest: opts.workerDigest,
APIKeyID: participants[0].APIKeyID,
UserID: participants[0].UserID,
Token: runToken,
EmulatorBaseURL: opts.emulatorBaseURL,
CallbackURL: opts.callbackURL,
CapacityProfile: "P24",
Config: map[string]any{
"workloads": []any{"gemini_image_edit", "multi_reference_video"},
"participants": participants,
"localClusterId": opts.clusterID,
"snapshotConfigHash": snapshotConfigHash,
"snapshotSha256": snapshotSHA,
},
})
if err != nil {
return err
}
if _, err := database.ActivateAcceptanceRun(ctx, run.ID); err != nil {
return err
}
outputPath, err := privateOutputPath(opts.output)
if err != nil {
return err
}
payload, err := json.MarshalIndent(runtimeFile{
SchemaVersion: "acceptance-runtime/v1",
LocalClusterID: opts.clusterID,
RunID: run.ID,
RunToken: runToken,
APIKeys: apiKeys,
Participants: participants,
GeminiModel: geminiModel,
VideoModel: videoModel,
EmulatorBaseURL: opts.emulatorBaseURL,
CallbackURL: opts.callbackURL,
ReleaseSHA: opts.releaseSHA,
APIImageDigest: opts.apiImageDigest,
WorkerImageDigest: opts.workerDigest,
SnapshotConfigHash: snapshotConfigHash,
SnapshotSHA256: snapshotSHA,
CreatedAt: time.Now().UTC(),
}, "", " ")
if err != nil {
return err
}
if err := os.WriteFile(outputPath, append(payload, '\n'), 0o600); err != nil {
return err
}
fmt.Printf(
"acceptance_bootstrap=PASS cluster_id=%s run_id=%s identities=%d config_hash=%s runtime_file=%s\n",
opts.clusterID,
run.ID,
len(participants),
snapshotConfigHash,
outputPath,
)
return nil
}
func verifyLocalDatabase(ctx context.Context, database *store.Store, clusterID string) error {
var marker string
err := database.Pool().QueryRow(ctx, `
SELECT COALESCE(value->>'clusterId', '')
FROM system_settings
WHERE setting_key = $1`, acceptancesnapshot.LocalClusterSettingKey).Scan(&marker)
if err != nil {
return fmt.Errorf("read local acceptance cluster marker: %w", err)
}
if marker != clusterID {
return errors.New("refusing bootstrap: local cluster marker mismatch")
}
mode, err := database.GetGatewayTrafficMode(ctx)
if err != nil {
return err
}
if mode.Mode != "live" {
return fmt.Errorf("refusing bootstrap while traffic mode is %s", mode.Mode)
}
return nil
}
func importedSnapshot(ctx context.Context, database *store.Store) (string, string, error) {
var configHash, snapshotSHA string
err := database.Pool().QueryRow(ctx, `
SELECT COALESCE(value->>'configHash', ''),
COALESCE(value->>'snapshotSha256', '')
FROM system_settings
WHERE setting_key = 'acceptance_snapshot'`).Scan(&configHash, &snapshotSHA)
if err != nil {
return "", "", fmt.Errorf("read imported acceptance snapshot: %w", err)
}
if len(configHash) != 64 || len(snapshotSHA) != 64 {
return "", "", errors.New("imported acceptance snapshot hashes are invalid")
}
return configHash, snapshotSHA, nil
}
func ensureAcceptanceIdentityDomain(ctx context.Context, database *store.Store) (string, string, error) {
var groupID string
err := database.Pool().QueryRow(ctx, `
INSERT INTO gateway_user_groups (
group_key, name, description, source, priority,
recharge_discount_policy, billing_discount_policy,
rate_limit_policy, quota_policy, metadata, status
)
VALUES (
'local-acceptance', 'Local Acceptance', 'Isolated local homologous acceptance identities',
'gateway', 1, '{"discountFactor":1}'::jsonb, '{"discountFactor":1}'::jsonb,
'{"rules":[]}'::jsonb, '{}'::jsonb,
'{"purpose":"local_acceptance","isolated":true}'::jsonb, 'active'
)
ON CONFLICT (group_key) DO UPDATE
SET name = EXCLUDED.name,
description = EXCLUDED.description,
priority = EXCLUDED.priority,
rate_limit_policy = EXCLUDED.rate_limit_policy,
quota_policy = EXCLUDED.quota_policy,
metadata = EXCLUDED.metadata,
status = 'active',
updated_at = now()
RETURNING id::text`).Scan(&groupID)
if err != nil {
return "", "", err
}
var tenantID string
err = database.Pool().QueryRow(ctx, `
INSERT INTO gateway_tenants (
tenant_key, source, external_tenant_id, name, default_user_group_id,
billing_profile, rate_limit_policy, auth_policy, metadata, status
)
VALUES (
'local-acceptance', 'gateway', 'local-acceptance', 'Local Acceptance',
$1::uuid, '{}'::jsonb, '{"rules":[]}'::jsonb, '{}'::jsonb,
'{"purpose":"local_acceptance","isolated":true}'::jsonb, 'active'
)
ON CONFLICT (tenant_key) DO UPDATE
SET default_user_group_id = EXCLUDED.default_user_group_id,
metadata = EXCLUDED.metadata,
status = 'active',
updated_at = now()
RETURNING id::text`, groupID).Scan(&tenantID)
return groupID, tenantID, err
}
func ensureAcceptanceUser(
ctx context.Context,
database *store.Store,
groupID string,
tenantID string,
ordinal int,
) (string, error) {
userKey := fmt.Sprintf("local-acceptance-%03d", ordinal)
roles := `["user"]`
if ordinal == 0 {
roles = `["manager"]`
}
var userID string
err := database.Pool().QueryRow(ctx, `
INSERT INTO gateway_users (
user_key, source, external_user_id, username, display_name,
gateway_tenant_id, tenant_id, tenant_key, default_user_group_id,
roles, auth_profile, metadata, status
)
VALUES (
$1, 'gateway', $1, $1, $2,
$3::uuid, 'local-acceptance', 'local-acceptance', $4::uuid,
$5::jsonb, '{}'::jsonb,
jsonb_build_object('purpose','local_acceptance','ordinal',$6,'isolated',true),
'active'
)
ON CONFLICT (user_key) DO UPDATE
SET gateway_tenant_id = EXCLUDED.gateway_tenant_id,
tenant_id = EXCLUDED.tenant_id,
tenant_key = EXCLUDED.tenant_key,
default_user_group_id = EXCLUDED.default_user_group_id,
roles = EXCLUDED.roles,
metadata = EXCLUDED.metadata,
status = 'active',
deleted_at = NULL,
updated_at = now()
RETURNING id::text`,
userKey,
fmt.Sprintf("Local Acceptance %03d", ordinal),
tenantID,
groupID,
roles,
ordinal,
).Scan(&userID)
return userID, err
}
func ensureAcceptanceAPIKey(
ctx context.Context,
database *store.Store,
userID string,
groupID string,
tenantID string,
ordinal int,
) (string, string, error) {
name := fmt.Sprintf("Local Acceptance %03d", ordinal)
var keyID, secret string
err := database.Pool().QueryRow(ctx, `
SELECT id::text, COALESCE(key_secret, '')
FROM gateway_api_keys
WHERE gateway_user_id = $1::uuid
AND name = $2
AND status = 'active'
AND deleted_at IS NULL
AND COALESCE(key_secret, '') <> ''
ORDER BY created_at
LIMIT 1`, userID, name).Scan(&keyID, &secret)
if err == nil {
return keyID, secret, nil
}
if !store.IsNotFound(err) {
return "", "", err
}
created, err := database.CreateAPIKey(ctx, store.CreateAPIKeyInput{
Name: name,
Scopes: []string{"image", "video"},
}, &auth.User{
ID: userID,
Username: fmt.Sprintf("local-acceptance-%03d", ordinal),
GatewayUserID: userID,
GatewayTenantID: tenantID,
TenantID: "local-acceptance",
TenantKey: "local-acceptance",
UserGroupID: groupID,
Source: "gateway",
})
if err != nil {
return "", "", err
}
if _, err := database.Pool().Exec(ctx, `
UPDATE gateway_api_keys
SET user_group_id = $2::uuid,
rate_limit_policy = '{"rules":[]}'::jsonb,
quota_policy = '{}'::jsonb,
updated_at = now()
WHERE id = $1::uuid`, created.APIKey.ID, groupID); err != nil {
return "", "", err
}
return created.APIKey.ID, created.Secret, nil
}
func ensureAcceptanceAccessRules(ctx context.Context, database *store.Store, groupID string) error {
_, err := database.Pool().Exec(ctx, `
WITH selected AS (
SELECT 'platform'::text AS resource_type, platform.id AS resource_id
FROM integration_platforms platform
WHERE COALESCE((platform.config->>'acceptanceSnapshot')::boolean, false)
UNION
SELECT 'platform_model', model.id
FROM platform_models model
JOIN integration_platforms platform ON platform.id = model.platform_id
WHERE COALESCE((platform.config->>'acceptanceSnapshot')::boolean, false)
UNION
SELECT 'base_model', model.base_model_id
FROM platform_models model
JOIN integration_platforms platform ON platform.id = model.platform_id
WHERE COALESCE((platform.config->>'acceptanceSnapshot')::boolean, false)
AND model.base_model_id IS NOT NULL
)
INSERT INTO gateway_access_rules (
subject_type, subject_id, resource_type, resource_id, effect,
priority, min_permission_level, conditions, metadata, status
)
SELECT 'user_group', $1::uuid, selected.resource_type, selected.resource_id,
'allow', 1, 0, '{}'::jsonb,
'{"purpose":"local_acceptance"}'::jsonb, 'active'
FROM selected
ON CONFLICT (subject_type, subject_id, resource_type, resource_id, effect)
DO UPDATE SET status = 'active', priority = 1, updated_at = now()`, groupID)
return err
}
func selectedModels(ctx context.Context, database *store.Store) (string, string, error) {
var gemini, video string
err := database.Pool().QueryRow(ctx, `
SELECT base_model.invocation_name
FROM platform_models model
JOIN integration_platforms platform ON platform.id = model.platform_id
JOIN base_model_catalog base_model ON base_model.id = model.base_model_id
WHERE platform.status = 'enabled'
AND COALESCE((platform.config->>'acceptanceSnapshot')::boolean, false)
AND model.enabled = true
AND model.model_type @> '["image_edit"]'::jsonb
ORDER BY platform.priority, model.created_at
LIMIT 1`).Scan(&gemini)
if err != nil {
return "", "", err
}
err = database.Pool().QueryRow(ctx, `
SELECT base_model.invocation_name
FROM platform_models model
JOIN integration_platforms platform ON platform.id = model.platform_id
JOIN base_model_catalog base_model ON base_model.id = model.base_model_id
WHERE platform.status = 'enabled'
AND COALESCE((platform.config->>'acceptanceSnapshot')::boolean, false)
AND model.enabled = true
AND model.model_type @> '["omni_video"]'::jsonb
ORDER BY platform.priority, model.created_at
LIMIT 1`).Scan(&video)
return gemini, video, err
}
func randomToken() (string, error) {
payload := make([]byte, 32)
if _, err := rand.Read(payload); err != nil {
return "", err
}
return hex.EncodeToString(payload), nil
}
func privateOutputPath(path string) (string, error) {
absolute, err := filepath.Abs(strings.TrimSpace(path))
if err != nil {
return "", err
}
if info, err := os.Lstat(absolute); err == nil {
if !info.Mode().IsRegular() || info.Mode()&os.ModeSymlink != 0 {
return "", errors.New("runtime output must be a regular non-symlink file")
}
} else if !os.IsNotExist(err) {
return "", err
}
if err := os.MkdirAll(filepath.Dir(absolute), 0o700); err != nil {
return "", err
}
return absolute, nil
}