mirror of
https://github.com/Rain-kl/OpenFlare.git
synced 2026-10-05 15:26:36 +08:00
[功能] 添加数据库观测数据清理功能，支持手动和自动清理策略
This commit is contained in:
ryan
@@ -0,0 +1,182 @@
|
||||
package service
|
||||
|
||||
import (
|
||||
"context"
|
||||
"errors"
|
||||
"fmt"
|
||||
"log/slog"
|
||||
"openflare/common"
|
||||
"openflare/model"
|
||||
"strings"
|
||||
"time"
|
||||
)
|
||||
|
||||
const (
|
||||
DatabaseCleanupTargetAccessLogs = "node_access_logs"
|
||||
DatabaseCleanupTargetMetricSnapshots = "node_metric_snapshots"
|
||||
DatabaseCleanupTargetRequestReports = "node_request_reports"
|
||||
)
|
||||
|
||||
var databaseCleanupTargets = map[string]string{
|
||||
DatabaseCleanupTargetAccessLogs: "访问日志",
|
||||
DatabaseCleanupTargetMetricSnapshots: "性能快照",
|
||||
DatabaseCleanupTargetRequestReports: "请求聚合",
|
||||
}
|
||||
|
||||
type DatabaseCleanupInput struct {
|
||||
Target string `json:"target"`
|
||||
RetentionDays *int `json:"retention_days"`
|
||||
}
|
||||
|
||||
type DatabaseCleanupResult struct {
|
||||
Target string `json:"target"`
|
||||
TargetLabel string `json:"target_label"`
|
||||
DeletedCount int64 `json:"deleted_count"`
|
||||
DeleteAll bool `json:"delete_all"`
|
||||
RetentionDays *int `json:"retention_days,omitempty"`
|
||||
Cutoff *time.Time `json:"cutoff,omitempty"`
|
||||
}
|
||||
|
||||
type DatabaseAutoCleanupSummary struct {
|
||||
RetentionDays int `json:"retention_days"`
|
||||
ExecutedAt time.Time `json:"executed_at"`
|
||||
Results []DatabaseCleanupResult `json:"results"`
|
||||
}
|
||||
|
||||
func CleanupDatabaseObservability(input DatabaseCleanupInput) (*DatabaseCleanupResult, error) {
|
||||
target := strings.TrimSpace(input.Target)
|
||||
targetLabel, ok := databaseCleanupTargets[target]
|
||||
if !ok {
|
||||
return nil, errors.New("unsupported cleanup target")
|
||||
}
|
||||
if input.RetentionDays != nil && *input.RetentionDays <= 0 {
|
||||
return nil, errors.New("retention_days 必须为大于 0 的整数")
|
||||
}
|
||||
|
||||
result := &DatabaseCleanupResult{
|
||||
Target: target,
|
||||
TargetLabel: targetLabel,
|
||||
DeleteAll: input.RetentionDays == nil,
|
||||
}
|
||||
|
||||
if input.RetentionDays == nil {
|
||||
deleted, err := deleteAllObservabilityRows(target)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
result.DeletedCount = deleted
|
||||
return result, nil
|
||||
}
|
||||
|
||||
retentionDays := *input.RetentionDays
|
||||
cutoff := time.Now().UTC().Add(-time.Duration(retentionDays) * 24 * time.Hour)
|
||||
deleted, err := deleteObservabilityRowsBefore(target, cutoff)
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
result.DeletedCount = deleted
|
||||
result.RetentionDays = &retentionDays
|
||||
result.Cutoff = &cutoff
|
||||
return result, nil
|
||||
}
|
||||
|
||||
func RunDatabaseAutoCleanupOnce(now time.Time) (*DatabaseAutoCleanupSummary, error) {
|
||||
if !common.DatabaseAutoCleanupEnabled {
|
||||
return nil, nil
|
||||
}
|
||||
if common.DatabaseAutoCleanupRetentionDays < 1 {
|
||||
return nil, fmt.Errorf("database auto cleanup retention_days must be at least 1")
|
||||
}
|
||||
|
||||
retentionDays := common.DatabaseAutoCleanupRetentionDays
|
||||
results := make([]DatabaseCleanupResult, 0, len(databaseCleanupTargets))
|
||||
for _, target := range []string{
|
||||
DatabaseCleanupTargetAccessLogs,
|
||||
DatabaseCleanupTargetMetricSnapshots,
|
||||
DatabaseCleanupTargetRequestReports,
|
||||
} {
|
||||
result, err := CleanupDatabaseObservability(DatabaseCleanupInput{
|
||||
Target: target,
|
||||
RetentionDays: &retentionDays,
|
||||
})
|
||||
if err != nil {
|
||||
return nil, err
|
||||
}
|
||||
results = append(results, *result)
|
||||
}
|
||||
|
||||
return &DatabaseAutoCleanupSummary{
|
||||
RetentionDays: retentionDays,
|
||||
ExecutedAt: now.UTC(),
|
||||
Results: results,
|
||||
}, nil
|
||||
}
|
||||
|
||||
func StartDatabaseAutoCleanupScheduler(ctx context.Context) {
|
||||
go func() {
|
||||
for {
|
||||
wait := time.Until(nextDatabaseAutoCleanupTime(time.Now()))
|
||||
timer := time.NewTimer(wait)
|
||||
select {
|
||||
case <-ctx.Done():
|
||||
timer.Stop()
|
||||
return
|
||||
case <-timer.C:
|
||||
}
|
||||
|
||||
summary, err := RunDatabaseAutoCleanupOnce(time.Now())
|
||||
if err != nil {
|
||||
slog.Error("database auto cleanup failed", "error", err)
|
||||
continue
|
||||
}
|
||||
if summary == nil {
|
||||
continue
|
||||
}
|
||||
totalDeleted := int64(0)
|
||||
for _, item := range summary.Results {
|
||||
totalDeleted += item.DeletedCount
|
||||
}
|
||||
slog.Info(
|
||||
"database auto cleanup completed",
|
||||
"retention_days",
|
||||
summary.RetentionDays,
|
||||
"deleted_count",
|
||||
totalDeleted,
|
||||
)
|
||||
}
|
||||
}()
|
||||
}
|
||||
|
||||
func nextDatabaseAutoCleanupTime(now time.Time) time.Time {
|
||||
next := time.Date(now.Year(), now.Month(), now.Day(), 3, 0, 0, 0, now.Location())
|
||||
if !next.After(now) {
|
||||
next = next.Add(24 * time.Hour)
|
||||
}
|
||||
return next
|
||||
}
|
||||
|
||||
func deleteAllObservabilityRows(target string) (int64, error) {
|
||||
switch target {
|
||||
case DatabaseCleanupTargetAccessLogs:
|
||||
return model.DeleteAllNodeAccessLogs(nil)
|
||||
case DatabaseCleanupTargetMetricSnapshots:
|
||||
return model.DeleteAllNodeMetricSnapshots(nil)
|
||||
case DatabaseCleanupTargetRequestReports:
|
||||
return model.DeleteAllNodeRequestReports(nil)
|
||||
default:
|
||||
return 0, errors.New("unsupported cleanup target")
|
||||
}
|
||||
}
|
||||
|
||||
func deleteObservabilityRowsBefore(target string, cutoff time.Time) (int64, error) {
|
||||
switch target {
|
||||
case DatabaseCleanupTargetAccessLogs:
|
||||
return model.DeleteNodeAccessLogsBefore(cutoff)
|
||||
case DatabaseCleanupTargetMetricSnapshots:
|
||||
return model.DeleteNodeMetricSnapshotsBefore(nil, cutoff)
|
||||
case DatabaseCleanupTargetRequestReports:
|
||||
return model.DeleteNodeRequestReportsBefore(nil, cutoff)
|
||||
default:
|
||||
return 0, errors.New("unsupported cleanup target")
|
||||
}
|
||||
}
|
||||
@@ -0,0 +1,165 @@
|
||||
package service
|
||||
|
||||
import (
|
||||
"openflare/common"
|
||||
"openflare/model"
|
||||
"testing"
|
||||
"time"
|
||||
)
|
||||
|
||||
func TestCleanupDatabaseObservabilityDeletesTargetedRows(t *testing.T) {
|
||||
setupServiceTestDB(t)
|
||||
|
||||
now := time.Now().UTC()
|
||||
if err := model.DB.Create(&model.NodeMetricSnapshot{
|
||||
NodeID: "node-a",
|
||||
CapturedAt: now.Add(-10 * 24 * time.Hour),
|
||||
CPUUsagePercent: 10,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed old metric snapshot: %v", err)
|
||||
}
|
||||
if err := model.DB.Create(&model.NodeMetricSnapshot{
|
||||
NodeID: "node-a",
|
||||
CapturedAt: now.Add(-12 * time.Hour),
|
||||
CPUUsagePercent: 20,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed recent metric snapshot: %v", err)
|
||||
}
|
||||
|
||||
retentionDays := 7
|
||||
result, err := CleanupDatabaseObservability(DatabaseCleanupInput{
|
||||
Target: DatabaseCleanupTargetMetricSnapshots,
|
||||
RetentionDays: &retentionDays,
|
||||
})
|
||||
if err != nil {
|
||||
t.Fatalf("CleanupDatabaseObservability failed: %v", err)
|
||||
}
|
||||
if result.DeleteAll {
|
||||
t.Fatal("expected retention cleanup instead of delete_all")
|
||||
}
|
||||
if result.DeletedCount != 1 {
|
||||
t.Fatalf("expected 1 deleted row, got %+v", result)
|
||||
}
|
||||
|
||||
rows, err := model.ListMetricSnapshotsSince(time.Time{})
|
||||
if err != nil {
|
||||
t.Fatalf("ListMetricSnapshotsSince failed: %v", err)
|
||||
}
|
||||
if len(rows) != 1 || rows[0].CPUUsagePercent != 20 {
|
||||
t.Fatalf("unexpected remaining metric snapshots: %+v", rows)
|
||||
}
|
||||
}
|
||||
|
||||
func TestCleanupDatabaseObservabilityDeletesAllRowsWhenRetentionMissing(t *testing.T) {
|
||||
setupServiceTestDB(t)
|
||||
|
||||
now := time.Now().UTC()
|
||||
if err := model.DB.Create(&model.NodeAccessLog{
|
||||
NodeID: "node-a",
|
||||
LoggedAt: now.Add(-3 * time.Hour),
|
||||
RemoteAddr: "203.0.113.1",
|
||||
Host: "example.com",
|
||||
Path: "/one",
|
||||
StatusCode: 200,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed first access log: %v", err)
|
||||
}
|
||||
if err := model.DB.Create(&model.NodeAccessLog{
|
||||
NodeID: "node-a",
|
||||
LoggedAt: now.Add(-2 * time.Hour),
|
||||
RemoteAddr: "203.0.113.2",
|
||||
Host: "example.com",
|
||||
Path: "/two",
|
||||
StatusCode: 502,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed second access log: %v", err)
|
||||
}
|
||||
|
||||
result, err := CleanupDatabaseObservability(DatabaseCleanupInput{
|
||||
Target: DatabaseCleanupTargetAccessLogs,
|
||||
})
|
||||
if err != nil {
|
||||
t.Fatalf("CleanupDatabaseObservability failed: %v", err)
|
||||
}
|
||||
if !result.DeleteAll || result.DeletedCount != 2 {
|
||||
t.Fatalf("unexpected delete-all result: %+v", result)
|
||||
}
|
||||
|
||||
rows, err := model.ListNodeAccessLogs(model.NodeAccessLogQuery{Page: 0, PageSize: 10})
|
||||
if err != nil {
|
||||
t.Fatalf("ListNodeAccessLogs failed: %v", err)
|
||||
}
|
||||
if len(rows) != 0 {
|
||||
t.Fatalf("expected all access logs deleted, got %+v", rows)
|
||||
}
|
||||
}
|
||||
|
||||
func TestRunDatabaseAutoCleanupOnceDeletesAllObservabilityTargets(t *testing.T) {
|
||||
setupServiceTestDB(t)
|
||||
|
||||
now := time.Now().UTC()
|
||||
if err := model.DB.Create(&model.NodeAccessLog{
|
||||
NodeID: "node-a",
|
||||
LoggedAt: now.Add(-48 * time.Hour),
|
||||
RemoteAddr: "203.0.113.10",
|
||||
Host: "example.com",
|
||||
Path: "/access",
|
||||
StatusCode: 200,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed access log: %v", err)
|
||||
}
|
||||
if err := model.DB.Create(&model.NodeMetricSnapshot{
|
||||
NodeID: "node-a",
|
||||
CapturedAt: now.Add(-48 * time.Hour),
|
||||
CPUUsagePercent: 10,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed metric snapshot: %v", err)
|
||||
}
|
||||
if err := model.DB.Create(&model.NodeRequestReport{
|
||||
NodeID: "node-a",
|
||||
WindowStartedAt: now.Add(-49 * time.Hour),
|
||||
WindowEndedAt: now.Add(-48 * time.Hour),
|
||||
RequestCount: 15,
|
||||
}).Error; err != nil {
|
||||
t.Fatalf("seed request report: %v", err)
|
||||
}
|
||||
|
||||
previousEnabled := common.DatabaseAutoCleanupEnabled
|
||||
previousRetentionDays := common.DatabaseAutoCleanupRetentionDays
|
||||
common.DatabaseAutoCleanupEnabled = true
|
||||
common.DatabaseAutoCleanupRetentionDays = 1
|
||||
t.Cleanup(func() {
|
||||
common.DatabaseAutoCleanupEnabled = previousEnabled
|
||||
common.DatabaseAutoCleanupRetentionDays = previousRetentionDays
|
||||
})
|
||||
|
||||
summary, err := RunDatabaseAutoCleanupOnce(now)
|
||||
if err != nil {
|
||||
t.Fatalf("RunDatabaseAutoCleanupOnce failed: %v", err)
|
||||
}
|
||||
if summary == nil || len(summary.Results) != 3 {
|
||||
t.Fatalf("unexpected auto cleanup summary: %+v", summary)
|
||||
}
|
||||
|
||||
accessLogs, err := model.ListNodeAccessLogs(model.NodeAccessLogQuery{Page: 0, PageSize: 10})
|
||||
if err != nil {
|
||||
t.Fatalf("ListNodeAccessLogs failed: %v", err)
|
||||
}
|
||||
if len(accessLogs) != 0 {
|
||||
t.Fatalf("expected auto cleanup to delete access logs, got %+v", accessLogs)
|
||||
}
|
||||
metricSnapshots, err := model.ListMetricSnapshotsSince(time.Time{})
|
||||
if err != nil {
|
||||
t.Fatalf("ListMetricSnapshotsSince failed: %v", err)
|
||||
}
|
||||
if len(metricSnapshots) != 0 {
|
||||
t.Fatalf("expected auto cleanup to delete metric snapshots, got %+v", metricSnapshots)
|
||||
}
|
||||
requestReports, err := model.ListRequestReportsSince(time.Time{})
|
||||
if err != nil {
|
||||
t.Fatalf("ListRequestReportsSince failed: %v", err)
|
||||
}
|
||||
if len(requestReports) != 0 {
|
||||
t.Fatalf("expected auto cleanup to delete request reports, got %+v", requestReports)
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user