Use full snapshot interval to compute Backup Ready conditions (#906)

* get full snapshot interval from schedule to determine backup ready condition

go.mod

@@ -13,6 +13,7 @@ require (
 	github.com/hashicorp/go-multierror v1.1.1
 	github.com/ironcore-dev/vgopath v0.1.4
 	github.com/prometheus/client_golang v1.18.0
+	github.com/robfig/cron/v3 v3.0.1
 	github.com/spf13/pflag v1.0.5
 	go.uber.org/zap v1.27.0
 	golang.org/x/exp v0.0.0-20240707233637-46b078467d37
@@ -111,7 +112,6 @@ require (
 	github.com/prometheus/client_model v0.6.0 // indirect
 	github.com/prometheus/common v0.45.0 // indirect
 	github.com/prometheus/procfs v0.12.0 // indirect
-	github.com/robfig/cron/v3 v3.0.1 // indirect
 	github.com/shopspring/decimal v1.3.1 // indirect
 	github.com/sirupsen/logrus v1.9.3 // indirect
 	github.com/spf13/afero v1.11.0 // indirect

internal/health/condition/check_backup_ready.go

@@ -10,6 +10,7 @@ import (
 	"time"
 
 	druidv1alpha1 "github.com/gardener/etcd-druid/api/v1alpha1"
+	"github.com/gardener/etcd-druid/internal/utils"
 
 	coordinationv1 "k8s.io/api/coordination/v1"
 	"k8s.io/apimachinery/pkg/types"
@@ -48,13 +49,21 @@ func (a *backupReadyCheck) Check(ctx context.Context, etcd druidv1alpha1.Etcd) R
 
 	//Fetch snapshot leases
 	var (
-		fullSnapErr, incrSnapErr error
-		fullSnapLease            = &coordinationv1.Lease{}
-		deltaSnapLease           = &coordinationv1.Lease{}
+		fullSnapErr, incrSnapErr, err error
+		fullSnapLease                 = &coordinationv1.Lease{}
+		fullSnapshotInterval          = 24 * time.Hour
+		deltaSnapLease                = &coordinationv1.Lease{}
 	)
 	fullSnapErr = a.cl.Get(ctx, types.NamespacedName{Name: getFullSnapLeaseName(&etcd), Namespace: etcd.ObjectMeta.Namespace}, fullSnapLease)
 	incrSnapErr = a.cl.Get(ctx, types.NamespacedName{Name: getDeltaSnapLeaseName(&etcd), Namespace: etcd.ObjectMeta.Namespace}, deltaSnapLease)
 
+	// Compute the full snapshot interval if full snapshot schedule is set
+	if etcd.Spec.Backup.FullSnapshotSchedule != nil {
+		if fullSnapshotInterval, err = utils.ComputeScheduleInterval(*etcd.Spec.Backup.FullSnapshotSchedule); err != nil {
+			return result
+		}
+	}
+
 	//Set status to Unknown if errors in fetching snapshot leases or lease never renewed
 	if fullSnapErr != nil || incrSnapErr != nil || (fullSnapLease.Spec.RenewTime == nil && deltaSnapLease.Spec.RenewTime == nil) {
 		return result
@@ -66,8 +75,8 @@ func (a *backupReadyCheck) Check(ctx context.Context, etcd druidv1alpha1.Etcd) R
 
 	if fullLeaseRenewTime == nil && deltaLeaseRenewTime != nil {
 		// Most probable during reconcile of existing clusters if fresh leases are created
-		// Treat backup as succeeded if delta snap lease renewal happens in the required time window and full snap lease is not older than 24h.
-		if time.Since(deltaLeaseRenewTime.Time) < 2*etcd.Spec.Backup.DeltaSnapshotPeriod.Duration && time.Since(fullLeaseCreateTime.Time) < 24*time.Hour {
+		// Treat backup as succeeded if delta snap lease renewal happens in the required time window and full snap lease is not older than fullSnapshotInterval
+		if time.Since(deltaLeaseRenewTime.Time) < 2*etcd.Spec.Backup.DeltaSnapshotPeriod.Duration && time.Since(fullLeaseCreateTime.Time) < fullSnapshotInterval {
 			result.reason = BackupSucceeded
 			result.message = "Delta snapshot backup succeeded"
 			result.status = druidv1alpha1.ConditionTrue
@@ -82,7 +91,7 @@ func (a *backupReadyCheck) Check(ctx context.Context, etcd druidv1alpha1.Etcd) R
 		}
 	} else if deltaLeaseRenewTime != nil && fullLeaseRenewTime != nil {
 		//Both snap leases are maintained. Both are expected to be renewed periodically
-		if time.Since(deltaLeaseRenewTime.Time) < 2*etcd.Spec.Backup.DeltaSnapshotPeriod.Duration && time.Since(fullLeaseRenewTime.Time) < 24*time.Hour {
+		if time.Since(deltaLeaseRenewTime.Time) < 2*etcd.Spec.Backup.DeltaSnapshotPeriod.Duration && time.Since(fullLeaseRenewTime.Time) < fullSnapshotInterval {
 			result.reason = BackupSucceeded
 			result.message = "Snapshot backup succeeded"
 			result.status = druidv1alpha1.ConditionTrue

internal/utils/miscellaneous.go

@@ -8,7 +8,9 @@ import (
 	"fmt"
 	"maps"
 	"strings"
+	"time"
 
+	"github.com/robfig/cron/v3"
 	"sigs.k8s.io/controller-runtime/pkg/client"
 )
 
@@ -63,3 +65,19 @@ func IfConditionOr[T any](condition bool, trueVal, falseVal T) T {
 	}
 	return falseVal
 }
+
+// ComputeScheduleInterval computes the interval between two activations for the given cron schedule.
+// Assumes that every cron activation is at equal intervals apart, based on cron schedules such as
+// "once every X hours", "once every Y days", "at 1:00pm on every Tuesday", etc.
+// TODO: write a new function to accurately compute the previous activation time from the cron schedule
+// in order to compute when the previous activation of the cron schedule was supposed to have occurred,
+// instead of relying on the assumption that all the cron activations are evenly spaced.
+func ComputeScheduleInterval(cronSchedule string) (time.Duration, error) {
+	schedule, err := cron.ParseStandard(cronSchedule)
+	if err != nil {
+		return 0, err
+	}
+	nextScheduledTime := schedule.Next(time.Now())
+	nextNextScheduledTime := schedule.Next(nextScheduledTime)
+	return nextNextScheduledTime.Sub(nextScheduledTime), nil
+}