cadvisor/info/v2/conversion.go
Rohit Agarwal 126fb2232e Add accelerator metrics to the API.
The structure is generic to support most hardware accelerators like
GPUs, TPUs etc.

Note that the prometheus label for id is called acc_id, so that it
doesn't conflict with some other label that maybe called id.
2017-11-01 14:41:35 -07:00

284 lines
8.2 KiB
Go

// Copyright 2015 Google Inc. All Rights Reserved.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
package v2
import (
"fmt"
"time"
"github.com/golang/glog"
"github.com/google/cadvisor/info/v1"
)
func machineFsStatsFromV1(fsStats []v1.FsStats) []MachineFsStats {
var result []MachineFsStats
for i := range fsStats {
stat := fsStats[i]
readDuration := time.Millisecond * time.Duration(stat.ReadTime)
writeDuration := time.Millisecond * time.Duration(stat.WriteTime)
ioDuration := time.Millisecond * time.Duration(stat.IoTime)
weightedDuration := time.Millisecond * time.Duration(stat.WeightedIoTime)
machineFsStat := MachineFsStats{
Device: stat.Device,
Type: stat.Type,
Capacity: &stat.Limit,
Usage: &stat.Usage,
Available: &stat.Available,
DiskStats: DiskStats{
ReadsCompleted: &stat.ReadsCompleted,
ReadsMerged: &stat.ReadsMerged,
SectorsRead: &stat.SectorsRead,
ReadDuration: &readDuration,
WritesCompleted: &stat.WritesCompleted,
WritesMerged: &stat.WritesMerged,
SectorsWritten: &stat.SectorsWritten,
WriteDuration: &writeDuration,
IoInProgress: &stat.IoInProgress,
IoDuration: &ioDuration,
WeightedIoDuration: &weightedDuration,
},
}
if stat.HasInodes {
machineFsStat.InodesFree = &stat.InodesFree
}
result = append(result, machineFsStat)
}
return result
}
func MachineStatsFromV1(cont *v1.ContainerInfo) []MachineStats {
var stats []MachineStats
var last *v1.ContainerStats
for i := range cont.Stats {
val := cont.Stats[i]
stat := MachineStats{
Timestamp: val.Timestamp,
}
if cont.Spec.HasCpu {
stat.Cpu = &val.Cpu
cpuInst, err := InstCpuStats(last, val)
if err != nil {
glog.Warningf("Could not get instant cpu stats: %v", err)
} else {
stat.CpuInst = cpuInst
}
last = val
}
if cont.Spec.HasMemory {
stat.Memory = &val.Memory
}
if cont.Spec.HasNetwork {
stat.Network = &NetworkStats{
// FIXME: Use reflection instead.
Tcp: TcpStat(val.Network.Tcp),
Tcp6: TcpStat(val.Network.Tcp6),
Interfaces: val.Network.Interfaces,
}
}
if cont.Spec.HasFilesystem {
stat.Filesystem = machineFsStatsFromV1(val.Filesystem)
}
// TODO(rjnagal): Handle load stats.
stats = append(stats, stat)
}
return stats
}
func ContainerStatsFromV1(containerName string, spec *v1.ContainerSpec, stats []*v1.ContainerStats) []*ContainerStats {
newStats := make([]*ContainerStats, 0, len(stats))
var last *v1.ContainerStats
for _, val := range stats {
stat := &ContainerStats{
Timestamp: val.Timestamp,
}
if spec.HasCpu {
stat.Cpu = &val.Cpu
cpuInst, err := InstCpuStats(last, val)
if err != nil {
glog.Warningf("Could not get instant cpu stats: %v", err)
} else {
stat.CpuInst = cpuInst
}
last = val
}
if spec.HasMemory {
stat.Memory = &val.Memory
}
if spec.HasNetwork {
// TODO: Handle TcpStats
stat.Network = &NetworkStats{
Tcp: TcpStat(val.Network.Tcp),
Tcp6: TcpStat(val.Network.Tcp6),
Interfaces: val.Network.Interfaces,
}
}
if spec.HasFilesystem {
if len(val.Filesystem) == 1 {
stat.Filesystem = &FilesystemStats{
TotalUsageBytes: &val.Filesystem[0].Usage,
BaseUsageBytes: &val.Filesystem[0].BaseUsage,
InodeUsage: &val.Filesystem[0].Inodes,
}
} else if len(val.Filesystem) > 1 && containerName != "/" {
// Cannot handle multiple devices per container.
glog.V(4).Infof("failed to handle multiple devices for container %s. Skipping Filesystem stats", containerName)
}
}
if spec.HasDiskIo {
stat.DiskIo = &val.DiskIo
}
if spec.HasCustomMetrics {
stat.CustomMetrics = val.CustomMetrics
}
if len(val.Accelerators) > 0 {
stat.Accelerators = val.Accelerators
}
// TODO(rjnagal): Handle load stats.
newStats = append(newStats, stat)
}
return newStats
}
func DeprecatedStatsFromV1(cont *v1.ContainerInfo) []DeprecatedContainerStats {
stats := make([]DeprecatedContainerStats, 0, len(cont.Stats))
var last *v1.ContainerStats
for _, val := range cont.Stats {
stat := DeprecatedContainerStats{
Timestamp: val.Timestamp,
HasCpu: cont.Spec.HasCpu,
HasMemory: cont.Spec.HasMemory,
HasNetwork: cont.Spec.HasNetwork,
HasFilesystem: cont.Spec.HasFilesystem,
HasDiskIo: cont.Spec.HasDiskIo,
HasCustomMetrics: cont.Spec.HasCustomMetrics,
}
if stat.HasCpu {
stat.Cpu = val.Cpu
cpuInst, err := InstCpuStats(last, val)
if err != nil {
glog.Warningf("Could not get instant cpu stats: %v", err)
} else {
stat.CpuInst = cpuInst
}
last = val
}
if stat.HasMemory {
stat.Memory = val.Memory
}
if stat.HasNetwork {
stat.Network.Interfaces = val.Network.Interfaces
}
if stat.HasFilesystem {
stat.Filesystem = val.Filesystem
}
if stat.HasDiskIo {
stat.DiskIo = val.DiskIo
}
if stat.HasCustomMetrics {
stat.CustomMetrics = val.CustomMetrics
}
// TODO(rjnagal): Handle load stats.
stats = append(stats, stat)
}
return stats
}
func InstCpuStats(last, cur *v1.ContainerStats) (*CpuInstStats, error) {
if last == nil {
return nil, nil
}
if !cur.Timestamp.After(last.Timestamp) {
return nil, fmt.Errorf("container stats move backwards in time")
}
if len(last.Cpu.Usage.PerCpu) != len(cur.Cpu.Usage.PerCpu) {
return nil, fmt.Errorf("different number of cpus")
}
timeDelta := cur.Timestamp.Sub(last.Timestamp)
if timeDelta <= 100*time.Millisecond {
return nil, fmt.Errorf("time delta unexpectedly small")
}
// Nanoseconds to gain precision and avoid having zero seconds if the
// difference between the timestamps is just under a second
timeDeltaNs := uint64(timeDelta.Nanoseconds())
convertToRate := func(lastValue, curValue uint64) (uint64, error) {
if curValue < lastValue {
return 0, fmt.Errorf("cumulative stats decrease")
}
valueDelta := curValue - lastValue
// Use float64 to keep precision
return uint64(float64(valueDelta) / float64(timeDeltaNs) * 1e9), nil
}
total, err := convertToRate(last.Cpu.Usage.Total, cur.Cpu.Usage.Total)
if err != nil {
return nil, err
}
percpu := make([]uint64, len(last.Cpu.Usage.PerCpu))
for i := range percpu {
var err error
percpu[i], err = convertToRate(last.Cpu.Usage.PerCpu[i], cur.Cpu.Usage.PerCpu[i])
if err != nil {
return nil, err
}
}
user, err := convertToRate(last.Cpu.Usage.User, cur.Cpu.Usage.User)
if err != nil {
return nil, err
}
system, err := convertToRate(last.Cpu.Usage.System, cur.Cpu.Usage.System)
if err != nil {
return nil, err
}
return &CpuInstStats{
Usage: CpuInstUsage{
Total: total,
PerCpu: percpu,
User: user,
System: system,
},
}, nil
}
// Get V2 container spec from v1 container info.
func ContainerSpecFromV1(specV1 *v1.ContainerSpec, aliases []string, namespace string) ContainerSpec {
specV2 := ContainerSpec{
CreationTime: specV1.CreationTime,
HasCpu: specV1.HasCpu,
HasMemory: specV1.HasMemory,
HasFilesystem: specV1.HasFilesystem,
HasNetwork: specV1.HasNetwork,
HasDiskIo: specV1.HasDiskIo,
HasCustomMetrics: specV1.HasCustomMetrics,
Image: specV1.Image,
Labels: specV1.Labels,
Envs: specV1.Envs,
}
if specV1.HasCpu {
specV2.Cpu.Limit = specV1.Cpu.Limit
specV2.Cpu.MaxLimit = specV1.Cpu.MaxLimit
specV2.Cpu.Mask = specV1.Cpu.Mask
}
if specV1.HasMemory {
specV2.Memory.Limit = specV1.Memory.Limit
specV2.Memory.Reservation = specV1.Memory.Reservation
specV2.Memory.SwapLimit = specV1.Memory.SwapLimit
}
if specV1.HasCustomMetrics {
specV2.CustomMetrics = specV1.CustomMetrics
}
specV2.Aliases = aliases
specV2.Namespace = namespace
return specV2
}