use istio client-go library instead of knative (#1661)

use istio client-go library instead of knative
bump kubernetes dependency version
change code coverage to codecov

vendor/k8s.io/client-go/util/workqueue/default_rate_limiters.go

@@ -35,7 +35,7 @@ type RateLimiter interface {
 }
 
 // DefaultControllerRateLimiter is a no-arg constructor for a default rate limiter for a workqueue.  It has
-// both overall and per-item rate limitting.  The overall is a token bucket and the per-item is exponential
+// both overall and per-item rate limiting.  The overall is a token bucket and the per-item is exponential
 func DefaultControllerRateLimiter() RateLimiter {
 	return NewMaxOfRateLimiter(
 		NewItemExponentialFailureRateLimiter(5*time.Millisecond, 1000*time.Second),
@@ -62,7 +62,7 @@ func (r *BucketRateLimiter) NumRequeues(item interface{}) int {
 func (r *BucketRateLimiter) Forget(item interface{}) {
 }
 
-// ItemExponentialFailureRateLimiter does a simple baseDelay*10^<num-failures> limit
+// ItemExponentialFailureRateLimiter does a simple baseDelay*2^<num-failures> limit
 // dealing with max failures and expiration are up to the caller
 type ItemExponentialFailureRateLimiter struct {
 	failuresLock sync.Mutex

vendor/k8s.io/client-go/util/workqueue/delaying_queue.go

@@ -18,6 +18,7 @@ package workqueue
 
 import (
 	"container/heap"
+	"sync"
 	"time"
 
 	"k8s.io/apimachinery/pkg/util/clock"
@@ -65,6 +66,8 @@ type delayingType struct {
 
 	// stopCh lets us signal a shutdown to the waiting loop
 	stopCh chan struct{}
+	// stopOnce guarantees we only signal shutdown a single time
+	stopOnce sync.Once
 
 	// heartbeat ensures we wait no more than maxWait before firing
 	heartbeat clock.Ticker
@@ -131,11 +134,14 @@ func (pq waitForPriorityQueue) Peek() interface{} {
 	return pq[0]
 }
 
-// ShutDown gives a way to shut off this queue
+// ShutDown stops the queue. After the queue drains, the returned shutdown bool
+// on Get() will be true. This method may be invoked more than once.
 func (q *delayingType) ShutDown() {
-	q.Interface.ShutDown()
-	close(q.stopCh)
-	q.heartbeat.Stop()
+	q.stopOnce.Do(func() {
+		q.Interface.ShutDown()
+		close(q.stopCh)
+		q.heartbeat.Stop()
+	})
 }
 
 // AddAfter adds the given item to the work queue after the given delay

vendor/k8s.io/client-go/util/workqueue/doc.go

@@ -23,4 +23,4 @@ limitations under the License.
 //  * Multiple consumers and producers. In particular, it is allowed for an
 //      item to be reenqueued while it is being processed.
 //  * Shutdown notifications.
-package workqueue
+package workqueue // import "k8s.io/client-go/util/workqueue"

vendor/k8s.io/client-go/util/workqueue/metrics.go

@@ -57,6 +57,11 @@ type SummaryMetric interface {
 	Observe(float64)
 }
 
+// HistogramMetric counts individual observations.
+type HistogramMetric interface {
+	Observe(float64)
+}
+
 type noopMetric struct{}
 
 func (noopMetric) Inc()            {}
@@ -73,9 +78,9 @@ type defaultQueueMetrics struct {
 	// total number of adds handled by a workqueue
 	adds CounterMetric
 	// how long an item stays in a workqueue
-	latency SummaryMetric
+	latency HistogramMetric
 	// how long processing an item from a workqueue takes
-	workDuration         SummaryMetric
+	workDuration         HistogramMetric
 	addTimes             map[t]time.Time
 	processingStartTimes map[t]time.Time
 
@@ -104,7 +109,7 @@ func (m *defaultQueueMetrics) get(item t) {
 	m.depth.Dec()
 	m.processingStartTimes[item] = m.clock.Now()
 	if startTime, exists := m.addTimes[item]; exists {
-		m.latency.Observe(m.sinceInMicroseconds(startTime))
+		m.latency.Observe(m.sinceInSeconds(startTime))
 		delete(m.addTimes, item)
 	}
 }
@@ -115,7 +120,7 @@ func (m *defaultQueueMetrics) done(item t) {
 	}
 
 	if startTime, exists := m.processingStartTimes[item]; exists {
-		m.workDuration.Observe(m.sinceInMicroseconds(startTime))
+		m.workDuration.Observe(m.sinceInSeconds(startTime))
 		delete(m.processingStartTimes, item)
 	}
 }
@@ -135,7 +140,7 @@ func (m *defaultQueueMetrics) updateUnfinishedWork() {
 	// Convert to seconds; microseconds is unhelpfully granular for this.
 	total /= 1000000
 	m.unfinishedWorkSeconds.Set(total)
-	m.longestRunningProcessor.Set(oldest) // in microseconds.
+	m.longestRunningProcessor.Set(oldest / 1000000)
 }
 
 type noMetrics struct{}
@@ -150,6 +155,11 @@ func (m *defaultQueueMetrics) sinceInMicroseconds(start time.Time) float64 {
 	return float64(m.clock.Since(start).Nanoseconds() / time.Microsecond.Nanoseconds())
 }
 
+// Gets the time since the specified start in seconds.
+func (m *defaultQueueMetrics) sinceInSeconds(start time.Time) float64 {
+	return m.clock.Since(start).Seconds()
+}
+
 type retryMetrics interface {
 	retry()
 }
@@ -170,10 +180,10 @@ func (m *defaultRetryMetrics) retry() {
 type MetricsProvider interface {
 	NewDepthMetric(name string) GaugeMetric
 	NewAddsMetric(name string) CounterMetric
-	NewLatencyMetric(name string) SummaryMetric
-	NewWorkDurationMetric(name string) SummaryMetric
+	NewLatencyMetric(name string) HistogramMetric
+	NewWorkDurationMetric(name string) HistogramMetric
 	NewUnfinishedWorkSecondsMetric(name string) SettableGaugeMetric
-	NewLongestRunningProcessorMicrosecondsMetric(name string) SettableGaugeMetric
+	NewLongestRunningProcessorSecondsMetric(name string) SettableGaugeMetric
 	NewRetriesMetric(name string) CounterMetric
 }
 
@@ -187,11 +197,11 @@ func (_ noopMetricsProvider) NewAddsMetric(name string) CounterMetric {
 	return noopMetric{}
 }
 
-func (_ noopMetricsProvider) NewLatencyMetric(name string) SummaryMetric {
+func (_ noopMetricsProvider) NewLatencyMetric(name string) HistogramMetric {
 	return noopMetric{}
 }
 
-func (_ noopMetricsProvider) NewWorkDurationMetric(name string) SummaryMetric {
+func (_ noopMetricsProvider) NewWorkDurationMetric(name string) HistogramMetric {
 	return noopMetric{}
 }
 
@@ -199,7 +209,7 @@ func (_ noopMetricsProvider) NewUnfinishedWorkSecondsMetric(name string) Settabl
 	return noopMetric{}
 }
 
-func (_ noopMetricsProvider) NewLongestRunningProcessorMicrosecondsMetric(name string) SettableGaugeMetric {
+func (_ noopMetricsProvider) NewLongestRunningProcessorSecondsMetric(name string) SettableGaugeMetric {
 	return noopMetric{}
 }
 
@@ -235,7 +245,7 @@ func (f *queueMetricsFactory) newQueueMetrics(name string, clock clock.Clock) qu
 		latency:                 mp.NewLatencyMetric(name),
 		workDuration:            mp.NewWorkDurationMetric(name),
 		unfinishedWorkSeconds:   mp.NewUnfinishedWorkSecondsMetric(name),
-		longestRunningProcessor: mp.NewLongestRunningProcessorMicrosecondsMetric(name),
+		longestRunningProcessor: mp.NewLongestRunningProcessorSecondsMetric(name),
 		addTimes:                map[t]time.Time{},
 		processingStartTimes:    map[t]time.Time{},
 	}

vendor/k8s.io/client-go/util/workqueue/parallelizer.go

@@ -25,14 +25,6 @@ import (
 
 type DoWorkPieceFunc func(piece int)
 
-// Parallelize is a very simple framework that allows for parallelizing
-// N independent pieces of work.
-//
-// Deprecated: Use ParallelizeUntil instead.
-func Parallelize(workers, pieces int, doWorkPiece DoWorkPieceFunc) {
-	ParallelizeUntil(nil, workers, pieces, doWorkPiece)
-}
-
 // ParallelizeUntil is a framework that allows for parallelizing N
 // independent pieces of work until done or the context is canceled.
 func ParallelizeUntil(ctx context.Context, workers, pieces int, doWorkPiece DoWorkPieceFunc) {