Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 6 additions & 0 deletions config/system-stats-monitor.json
Original file line number Diff line number Diff line change
Expand Up @@ -94,6 +94,12 @@
},
"memory/percent_used": {
"displayName": "memory/percent_used"
},
"memory/zswap_bytes_used": {
"displayName": "memory/zswap_bytes_used"
},
"memory/zswap_compression_efficiency": {
"displayName": "memory/zswap_compression_efficiency"
}
}
},
Expand Down
2 changes: 2 additions & 0 deletions pkg/exporters/stackdriver/metric_conversion_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -67,6 +67,8 @@ func TestNPDMetricToSDMetric(t *testing.T) {
{metrics.MemoryPageCacheUsedID, "compute.googleapis.com/guest/memory/page_cache_used", "compute.googleapis.com"},
{metrics.MemoryUnevictableUsedID, "compute.googleapis.com/guest/memory/unevictable_used", "compute.googleapis.com"},
{metrics.MemoryPercentUsedID, "compute.googleapis.com/guest/memory/percent_used", "compute.googleapis.com"},
{metrics.MemoryZswapBytesUsedID, "compute.googleapis.com/guest/memory/zswap/bytes_used", "compute.googleapis.com"},
{metrics.MemoryZswapCompressionEfficiencyID, "compute.googleapis.com/guest/memory/zswap/compression_efficiency", "compute.googleapis.com"},

// Problem metrics
{metrics.ProblemCounterID, "compute.googleapis.com/guest/system/problem_count", "compute.googleapis.com"},
Expand Down
92 changes: 47 additions & 45 deletions pkg/exporters/stackdriver/stackdriver_exporter.go
Original file line number Diff line number Diff line change
Expand Up @@ -57,51 +57,53 @@ func init() {
const exporterName = "stackdriver"

var NPDMetricToSDMetric = map[metrics.MetricID]string{
metrics.CPURunnableTaskCountID: "compute.googleapis.com/guest/cpu/runnable_task_count",
metrics.CPUUsageTimeID: "compute.googleapis.com/guest/cpu/usage_time",
metrics.CPULoad1m: "compute.googleapis.com/guest/cpu/load_1m",
metrics.CPULoad5m: "compute.googleapis.com/guest/cpu/load_5m",
metrics.CPULoad15m: "compute.googleapis.com/guest/cpu/load_15m",
metrics.DiskAvgQueueLenID: "compute.googleapis.com/guest/disk/queue_length",
metrics.DiskBytesUsedID: "compute.googleapis.com/guest/disk/bytes_used",
metrics.DiskPercentUsedID: "compute.googleapis.com/guest/disk/percent_used",
metrics.DiskIOTimeID: "compute.googleapis.com/guest/disk/io_time",
metrics.DiskMergedOpsCountID: "compute.googleapis.com/guest/disk/merged_operation_count",
metrics.DiskOpsBytesID: "compute.googleapis.com/guest/disk/operation_bytes_count",
metrics.DiskOpsCountID: "compute.googleapis.com/guest/disk/operation_count",
metrics.DiskOpsTimeID: "compute.googleapis.com/guest/disk/operation_time",
metrics.DiskWeightedIOID: "compute.googleapis.com/guest/disk/weighted_io_time",
metrics.HostUptimeID: "compute.googleapis.com/guest/system/uptime",
metrics.MemoryAnonymousUsedID: "compute.googleapis.com/guest/memory/anonymous_used",
metrics.MemoryBytesUsedID: "compute.googleapis.com/guest/memory/bytes_used",
metrics.MemoryDirtyUsedID: "compute.googleapis.com/guest/memory/dirty_used",
metrics.MemoryPageCacheUsedID: "compute.googleapis.com/guest/memory/page_cache_used",
metrics.MemoryUnevictableUsedID: "compute.googleapis.com/guest/memory/unevictable_used",
metrics.MemoryPercentUsedID: "compute.googleapis.com/guest/memory/percent_used",
metrics.ProblemCounterID: "compute.googleapis.com/guest/system/problem_count",
metrics.ProblemGaugeID: "compute.googleapis.com/guest/system/problem_state",
metrics.OSFeatureID: "compute.googleapis.com/guest/system/os_feature_enabled",
metrics.SystemProcessesTotal: "kubernetes.io/internal/node/guest/system/processes_total",
metrics.SystemProcsRunning: "kubernetes.io/internal/node/guest/system/procs_running",
metrics.SystemProcsBlocked: "kubernetes.io/internal/node/guest/system/procs_blocked",
metrics.SystemInterruptsTotal: "kubernetes.io/internal/node/guest/system/interrupts_total",
metrics.SystemCPUStat: "kubernetes.io/internal/node/guest/system/cpu_stat",
metrics.NetDevRxBytes: "kubernetes.io/internal/node/guest/net/rx_bytes",
metrics.NetDevRxPackets: "kubernetes.io/internal/node/guest/net/rx_packets",
metrics.NetDevRxErrors: "kubernetes.io/internal/node/guest/net/rx_errors",
metrics.NetDevRxDropped: "kubernetes.io/internal/node/guest/net/rx_dropped",
metrics.NetDevRxFifo: "kubernetes.io/internal/node/guest/net/rx_fifo",
metrics.NetDevRxFrame: "kubernetes.io/internal/node/guest/net/rx_frame",
metrics.NetDevRxCompressed: "kubernetes.io/internal/node/guest/net/rx_compressed",
metrics.NetDevRxMulticast: "kubernetes.io/internal/node/guest/net/rx_multicast",
metrics.NetDevTxBytes: "kubernetes.io/internal/node/guest/net/tx_bytes",
metrics.NetDevTxPackets: "kubernetes.io/internal/node/guest/net/tx_packets",
metrics.NetDevTxErrors: "kubernetes.io/internal/node/guest/net/tx_errors",
metrics.NetDevTxDropped: "kubernetes.io/internal/node/guest/net/tx_dropped",
metrics.NetDevTxFifo: "kubernetes.io/internal/node/guest/net/tx_fifo",
metrics.NetDevTxCollisions: "kubernetes.io/internal/node/guest/net/tx_collisions",
metrics.NetDevTxCarrier: "kubernetes.io/internal/node/guest/net/tx_carrier",
metrics.NetDevTxCompressed: "kubernetes.io/internal/node/guest/net/tx_compressed",
metrics.CPURunnableTaskCountID: "compute.googleapis.com/guest/cpu/runnable_task_count",
metrics.CPUUsageTimeID: "compute.googleapis.com/guest/cpu/usage_time",
metrics.CPULoad1m: "compute.googleapis.com/guest/cpu/load_1m",
metrics.CPULoad5m: "compute.googleapis.com/guest/cpu/load_5m",
metrics.CPULoad15m: "compute.googleapis.com/guest/cpu/load_15m",
metrics.DiskAvgQueueLenID: "compute.googleapis.com/guest/disk/queue_length",
metrics.DiskBytesUsedID: "compute.googleapis.com/guest/disk/bytes_used",
metrics.DiskPercentUsedID: "compute.googleapis.com/guest/disk/percent_used",
metrics.DiskIOTimeID: "compute.googleapis.com/guest/disk/io_time",
metrics.DiskMergedOpsCountID: "compute.googleapis.com/guest/disk/merged_operation_count",
metrics.DiskOpsBytesID: "compute.googleapis.com/guest/disk/operation_bytes_count",
metrics.DiskOpsCountID: "compute.googleapis.com/guest/disk/operation_count",
metrics.DiskOpsTimeID: "compute.googleapis.com/guest/disk/operation_time",
metrics.DiskWeightedIOID: "compute.googleapis.com/guest/disk/weighted_io_time",
metrics.HostUptimeID: "compute.googleapis.com/guest/system/uptime",
metrics.MemoryAnonymousUsedID: "compute.googleapis.com/guest/memory/anonymous_used",
metrics.MemoryBytesUsedID: "compute.googleapis.com/guest/memory/bytes_used",
metrics.MemoryDirtyUsedID: "compute.googleapis.com/guest/memory/dirty_used",
metrics.MemoryPageCacheUsedID: "compute.googleapis.com/guest/memory/page_cache_used",
metrics.MemoryUnevictableUsedID: "compute.googleapis.com/guest/memory/unevictable_used",
metrics.MemoryPercentUsedID: "compute.googleapis.com/guest/memory/percent_used",
metrics.MemoryZswapBytesUsedID: "compute.googleapis.com/guest/memory/zswap/bytes_used",
metrics.MemoryZswapCompressionEfficiencyID: "compute.googleapis.com/guest/memory/zswap/compression_efficiency",
Comment on lines +81 to +82

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Is this metric registered and/or launched? I cannot find any documentation and some issues came up when testing locally.

  1. CreateTimeSeries rejected on both spellings:
control  guest/memory/percent_used            → {}  (accepted)
PR type  guest/memory/zswap/bytes_used        → 400 INVALID_ARGUMENT "unknown metric type"
flat     guest/memory/zswap_bytes_used        → 400 INVALID_ARGUMENT "unknown metric type"
  1. Not in the API descriptors:
$ GET /v3/projects/…/metricDescriptors?filter=starts_with("compute.googleapis.com/guest")
-- No match for zswap
  1. Running on stackdriver fails on every export:
… metric.type="compute.googleapis.com/guest/memory/zswap/compression_efficiency" … unknown metric type
… metric.type="compute.googleapis.com/guest/memory/zswap/bytes_used" … unknown metric type

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Hi, thanks for the catch! I updated the documentation, the reason for the error is because the metric descriptors have not yet been registered in the backend, once that happens, the errors should be resolved.

metrics.ProblemCounterID: "compute.googleapis.com/guest/system/problem_count",
metrics.ProblemGaugeID: "compute.googleapis.com/guest/system/problem_state",
metrics.OSFeatureID: "compute.googleapis.com/guest/system/os_feature_enabled",
metrics.SystemProcessesTotal: "kubernetes.io/internal/node/guest/system/processes_total",
metrics.SystemProcsRunning: "kubernetes.io/internal/node/guest/system/procs_running",
metrics.SystemProcsBlocked: "kubernetes.io/internal/node/guest/system/procs_blocked",
metrics.SystemInterruptsTotal: "kubernetes.io/internal/node/guest/system/interrupts_total",
metrics.SystemCPUStat: "kubernetes.io/internal/node/guest/system/cpu_stat",
metrics.NetDevRxBytes: "kubernetes.io/internal/node/guest/net/rx_bytes",
metrics.NetDevRxPackets: "kubernetes.io/internal/node/guest/net/rx_packets",
metrics.NetDevRxErrors: "kubernetes.io/internal/node/guest/net/rx_errors",
metrics.NetDevRxDropped: "kubernetes.io/internal/node/guest/net/rx_dropped",
metrics.NetDevRxFifo: "kubernetes.io/internal/node/guest/net/rx_fifo",
metrics.NetDevRxFrame: "kubernetes.io/internal/node/guest/net/rx_frame",
metrics.NetDevRxCompressed: "kubernetes.io/internal/node/guest/net/rx_compressed",
metrics.NetDevRxMulticast: "kubernetes.io/internal/node/guest/net/rx_multicast",
metrics.NetDevTxBytes: "kubernetes.io/internal/node/guest/net/tx_bytes",
metrics.NetDevTxPackets: "kubernetes.io/internal/node/guest/net/tx_packets",
metrics.NetDevTxErrors: "kubernetes.io/internal/node/guest/net/tx_errors",
metrics.NetDevTxDropped: "kubernetes.io/internal/node/guest/net/tx_dropped",
metrics.NetDevTxFifo: "kubernetes.io/internal/node/guest/net/tx_fifo",
metrics.NetDevTxCollisions: "kubernetes.io/internal/node/guest/net/tx_collisions",
metrics.NetDevTxCarrier: "kubernetes.io/internal/node/guest/net/tx_carrier",
metrics.NetDevTxCompressed: "kubernetes.io/internal/node/guest/net/tx_compressed",
}

func getMetricTypeConversionFunction(customMetricPrefix string) func(string) string {
Expand Down
2 changes: 2 additions & 0 deletions pkg/systemstatsmonitor/README.md
Original file line number Diff line number Diff line change
Expand Up @@ -77,6 +77,8 @@ Below metrics are collected from `memory` component:
* `memory_page_cache_used`: Page cache memory usage, in Bytes. Memory usage state is reported under the `state` metric label (e.g. `active`, `inactive`). `active` means the memory has been used more recently and usually not reclaimed until needed. Summing values of all states yields the total page cache memory used.
* `memory_unevictable_used`: [Unevictable memory][/proc doc] usage, in Bytes.
* `memory_dirty_used`: Dirty pages usage, in Bytes. Memory usage state is reported under the `state` metric label (e.g. `dirty`, `writeback`). `dirty` means the memory is waiting to be written back to disk, and `writeback` means the memory is actively being written back to disk.
* `memory_zswap_bytes_used`: Zswap memory usage, in Bytes. Collected from `/proc/meminfo` (`Zswap`).
* `memory_zswap_compression_efficiency`: Zswap compression efficiency ratio (`Zswapped` / `Zswap`).

### OS features

Expand Down
39 changes: 33 additions & 6 deletions pkg/systemstatsmonitor/memory_collector.go
Original file line number Diff line number Diff line change
Expand Up @@ -24,12 +24,14 @@ import (
)

type memoryCollector struct {
mBytesUsed *metrics.Int64Metric
mPercentUsed *metrics.Float64Metric
mAnonymousUsed *metrics.Int64Metric
mPageCacheUsed *metrics.Int64Metric
mUnevictableUsed *metrics.Int64Metric
mDirtyUsed *metrics.Int64Metric
mBytesUsed *metrics.Int64Metric
mPercentUsed *metrics.Float64Metric
mAnonymousUsed *metrics.Int64Metric
mPageCacheUsed *metrics.Int64Metric
mUnevictableUsed *metrics.Int64Metric
mDirtyUsed *metrics.Int64Metric
mZswapBytesUsed *metrics.Int64Metric
mZswapCompressionEfficiency *metrics.Float64Metric

config *ssmtypes.MemoryStatsConfig
}
Expand Down Expand Up @@ -105,5 +107,30 @@ func NewMemoryCollectorOrDie(memoryConfig *ssmtypes.MemoryStatsConfig) *memoryCo
klog.Fatalf("Error initializing metric for %q: %v", metrics.MemoryDirtyUsedID, err)
}

if _, ok := memoryConfig.MetricsConfigs[string(metrics.MemoryZswapBytesUsedID)]; ok {

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Can you add some unit test?

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

and can you squash the commits after the update?

Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I added unit tests for zswap to the commit and squashed it all into one commit, thanks

mc.mZswapBytesUsed, err = metrics.NewInt64Metric(
metrics.MemoryZswapBytesUsedID,
memoryConfig.MetricsConfigs[string(metrics.MemoryZswapBytesUsedID)].DisplayName,
"Zswap usage in bytes",
"Byte",
metrics.LastValue,
[]string{})
if err != nil {
klog.Fatalf("Error initializing metric for %q: %v", metrics.MemoryZswapBytesUsedID, err)
}
}
if _, ok := memoryConfig.MetricsConfigs[string(metrics.MemoryZswapCompressionEfficiencyID)]; ok {
mc.mZswapCompressionEfficiency, err = metrics.NewFloat64Metric(
metrics.MemoryZswapCompressionEfficiencyID,
memoryConfig.MetricsConfigs[string(metrics.MemoryZswapCompressionEfficiencyID)].DisplayName,
"Zswap compression efficiency ratio (uncompressed/compressed)",
"1",
metrics.LastValue,
[]string{})
if err != nil {
klog.Fatalf("Error initializing metric for %q: %v", metrics.MemoryZswapCompressionEfficiencyID, err)
}
}

return &mc
}
38 changes: 38 additions & 0 deletions pkg/systemstatsmonitor/memory_collector_test.go
Original file line number Diff line number Diff line change
Expand Up @@ -17,12 +17,50 @@ limitations under the License.
package systemstatsmonitor

import (
"encoding/json"
"testing"

"github.com/stretchr/testify/assert"

ssmtypes "k8s.io/node-problem-detector/pkg/systemstatsmonitor/types"
)

const (
fakeZswapConfig = `
{
"metricsConfigs": {
"memory/zswap_bytes_used": {
"displayName": "memory/zswap_bytes_used"
},
"memory/zswap_compression_efficiency": {
"displayName": "memory/zswap_compression_efficiency"
}
}
}
`
)

func TestMemoryCollector(t *testing.T) {
// Original test
mc := NewMemoryCollectorOrDie(&ssmtypes.MemoryStatsConfig{})
mc.collect()

// Ensure zswap metrics are nil when not in config
assert.Nil(t, mc.mZswapBytesUsed)
assert.Nil(t, mc.mZswapCompressionEfficiency)
}

func TestMemoryCollectorZswap(t *testing.T) {
cfg := &ssmtypes.MemoryStatsConfig{}
if err := json.Unmarshal([]byte(fakeZswapConfig), cfg); err != nil {
t.Fatalf("cannot load memory config: %s", err)
}

mc := NewMemoryCollectorOrDie(cfg)

// Ensure zswap metrics are initialized when in config
assert.NotNil(t, mc.mZswapBytesUsed)
assert.NotNil(t, mc.mZswapCompressionEfficiency)

mc.collect()
}
20 changes: 20 additions & 0 deletions pkg/systemstatsmonitor/memory_collector_unix.go
Original file line number Diff line number Diff line change
Expand Up @@ -120,4 +120,24 @@ func (mc *memoryCollector) collect() {
klog.Errorf("Failed to record unevictable used memory: %v", err)
}
}

if mc.mZswapBytesUsed != nil {
var bytesUsed int64
if meminfo.Zswap != nil && *meminfo.Zswap > 0 {
bytesUsed = int64(*meminfo.Zswap) * 1024 // kB -> B
}
if err := mc.mZswapBytesUsed.Record(map[string]string{}, bytesUsed); err != nil {
klog.Errorf("Failed to record zswap bytes used: %v", err)
}
}

if mc.mZswapCompressionEfficiency != nil {
var efficiency float64
if meminfo.Zswap != nil && *meminfo.Zswap > 0 && meminfo.Zswapped != nil {
efficiency = float64(*meminfo.Zswapped) / float64(*meminfo.Zswap)
}
if err := mc.mZswapCompressionEfficiency.Record(map[string]string{}, efficiency); err != nil {
klog.Errorf("Failed to record zswap compression efficiency: %v", err)
}
}
}
Loading