types

type ActionConfig struct {
	Name    string          `json:"name"`
	ArgsStr json.RawMessage `json:"args"`
	Args    interface{}     `json:"-"`
}

type AggregationsConfig struct {
	// MemoryAggregationInterval is the length of a single interval, for
	// which the peak memory usage is computed.
	// Memory usage peaks are aggregated in multiples of this interval. In other words
	// there is one memory usage sample per interval (the maximum usage over that
	// interval).
	MemoryAggregationInterval times.Duration `json:"memory_aggregation_interval"`
	// MemoryAggregationWindowIntervalCount is the number of consecutive MemoryAggregationIntervals
	// which make up the MemoryAggregationWindowLength which in turn is the period for memory
	// usage aggregation by VPA.
	MemoryAggregationIntervalCount int64 `json:"memory_aggregation_interval_count"`
	// MemoryHistogramDecayHalfLife is the amount of time it takes a historical
	// memory usage sample to lose half of its weight. In other words, a fresh
	// usage sample is twice as 'important' as one with age equal to the half
	// life period.
	MemoryHistogramDecayHalfLife times.Duration `json:"memory_histogram_decay_half_life"`
	// CPUHistogramDecayHalfLife is the amount of time it takes a historical
	// CPU usage sample to lose half of its weight.
	CPUHistogramDecayHalfLife times.Duration `json:"cpu_histogram_decay_half_life"`
}

type AlarmChannel struct {
	LocalAlarm  *LocalAlarm  `json:"local"`
	RemoteAlarm *RemoteAlarm `json:"remote"`
}

type AlarmConfig struct {
	Enable                 bool           `json:"enable"`
	Cluster                string         `json:"cluster"`
	MessageBatch           int            `json:"message_batch"`
	MessageDelay           times.Duration `json:"message_delay"`
	ChannelName            string         `json:"channel_name"`
	IgnoreAlarmWhenSilence bool           `json:"ignore_alarm_when_silence"`
	AlarmChannel           `json:"alarm_channel"`
}

type CPIManagerConfig struct {
	// I want this feature disabled by default
	Enable            bool           `json:"enable"`
	WindowDuration    times.Duration `json:"window_duration"`
	PrometheusAddrStr string         `json:"prometheus_addr"`
	PrometheusAddr    url.URL        `json:"-"`
	MaxJobSpecRange   times.Duration `json:"max_job_spec_range"`
}

type CaelusConfig struct {
	K8sConfig    K8sConfig          `json:"k8s_config"`
	CheckPoint   CheckPointConfig   `json:"check_point"`
	TaskType     TaskTypeConfig     `json:"task_type"`
	NodeResource NodeResourceConfig `json:"node_resource"`
	// If multiple predicts, the first one is used for real prediction. The left are experiment predicts, caelus will
	// only feeds samples to them and expose predict metrics for them.
	Predicts        []PredictConfig       `json:"predicts"`
	Metrics         MetricsCollectConfig  `json:"metrics"`
	ResourceIsolate ResourceIsolateConfig `json:"resource_isolate"`
	CpiManager      CPIManagerConfig      `json:"cpi_manager"`
	Alarm           AlarmConfig           `json:"alarm"`
	Online          OnlineConfig          `json:"online"`
	DiskQuota       DiskQuotaConfig       `json:"disk_quota"`
}

type CheckPointConfig struct {
	CheckPointDir   string `json:"check_point_dir"`
	NodeResourceKey string `json:"node_resource_key"`
}

type ComponentConfig struct {
	Cgroup  string `json:"cgroup"`
	Command string `json:"command"`
}

type CpuIsolateConfig struct {
	// AutoDetect will enable bt feature if supported, and quota as the second choice.
	AutoDetect bool `json:"auto_detect"`
	// ManagePolicy assigns cpu manage policy
	ManagePolicy   string         `json:"manage_policy"`
	CpuSetConfig   CpuSetConfig   `json:"cpuset_config"`
	CpuQuotaConfig CpuQuotaConfig `json:"cpu_quota_config"`
	// KubeletStatic check if cpu manager policy for kubelet is static
	KubeletStatic bool `json:"-"`
}

type CpuQuotaConfig struct {
	// set offline job weights, just for quota policy
	OfflineShare *uint64 `json:"offline_share"`
}

type CpuSetConfig struct {
	// isolate online jobs with offline jobs
	EnableOnlineIsolate bool `json:"enable_online_isolate"`
	// cpu list, which offline job will not be assigned
	ReservedCpus string `json:"reserved_cpus"`
}

type CustomMetric struct {
	MetricServerAddr string         `json:"metric_server_addr"`
	CollectInterval  times.Duration `json:"collect_interval"`
}

type DetectActionConfig struct {
	Detects []*DetectConfig `json:"detects"`
	Actions []*ActionConfig `json:"actions"`
}

type DetectConfig struct {
	Name    string          `json:"name"`
	ArgsStr json.RawMessage `json:"args"`
	Args    interface{}     `json:"-"`
}

type Devices struct {
	// Ifaces are the network interfaces, e.g. eth0, those not exist or down will be filter out
	// these ifaces will be assigned to metrics.node.ifaces
	IfacesWithProperty []string `json:"ifaces_xxx"`
	Ifaces             []string `json:"-"`
	// DiskNames are the disk names, e.g. sda, vda, those not exist will be filter out
	// these ifaces will be assigned to metrics.node.deviceNames
	DiskNames []string `json:"disk_names"`
}

type DiskQuotaConfig struct {
	Enabled     bool           `json:"enabled"`
	CheckPeriod times.Duration `json:"check_period"`
	// such as docker or containerd
	ContainerRuntime string `json:"container_runtime"`
	// quota size just for offline job, online jobs need to announce in annotations
	VolumeSizes map[VolumeType]*DiskQuotaSize `json:"volume_sizes"`
}

type DiskQuotaSize struct {
	Quota      uint64 `json:"quota"`
	Inodes     uint64 `json:"inodes"`
	QuotaUsed  uint64 `json:"-"`
	InodesUsed uint64 `json:"-"`
}

type EWMAArgs struct {
	Metric string `json:"metric"`
	Nr     int    `json:"nr"`
}

type ExpressionArgs struct {
	Expression      string         `json:"expression"`
	WarningCount    int            `json:"warning_count"`
	WarningDuration times.Duration `json:"warning_duration"`
}

type HealthCheckConfig struct {
	Disable      bool         `json:"disable"`
	RuleNodes    []string     `json:"rule_nodes"`
	RuleCheck    RuleCheck    `json:"rule_check"`
	CgroupNotify NotifyConfig `json:"cgroup_notify"`
	// assign the value when initialize
	PredictReserved *Resource `json:"-"`
}

type K8sConfig struct {
	KubeletRootDir string `json:"kubelet_root_dir"`
}

type LocalAlarm struct {
	Executor string `json:"executor"`
}

type LocalPredictConfig struct {
	// Minimum CPU recommendation for a pod
	PodMinCPUMillicores float64 `json:"pod_min_cpu_millicores"`
	// Minimum memory recommendation for a pod
	PodMinMemoryMb float64 `json:"pod_min_memory_mb"`
	// Fraction of usage added as the safety margin to the recommended request
	SafetyMarginFraction float64 `json:"safety_margin_fraction"`
	// cpu usage percentile to recommend cpu resource
	CPUPercentile float64 `json:"cpu_percentile"`
	// memory usage percentile to recommend cpu resource
	MemoryPeaksPercentile float64 `json:"memory_peaks_percentile"`
	// AggregationsConfig is used to configure aggregation behaviour.
	AggregationsConfig `json:",inline"`
	// Enable tune cpu weight if cpu usage is anomaly
	EnableTuneCPUWeight bool `json:"enable_tune_cpu_weight"`
	// AnomalyDetectorMovingWindow defines how long the moving window of anomaly detector should keep
	AnomalyDetectorMovingWindow times.Duration `json:"anomaly_detector_moving_window"`
	// If detect cpu usage increasing anomaly, the weight of the anomaly sample
	// Base weight is 100
	IncreasingAnomalyWeightFactor int64 `json:"increasing_anomaly_weight_factor"`
	// If detect cpu usage decreasing anomaly, the weight of the anomaly sample
	// Base weight is 100
	DecreasingAnomalyWeightFactor int64 `json:"decreasing_anomaly_weight_factor"`
}

type MemoryNotifyConfig struct {
	Pressures []MemoryPressureNotifyConfig `json:"pressures"`
	Usages    []MemoryUsageNotifyConfig    `json:"usages"`
}

type MemoryPressureNotifyConfig struct {
	Cgroups       []string `json:"cgroups"`
	PressureLevel string   `json:"pressure_level"`
	// assign time duration the pressure has kept
	Duration times.Duration `json:"duration"`
	// assign event number in the duration time
	Count int `json:"count"`
}

type MemoryUsageNotifyConfig struct {
	Cgroups []string `json:"cgroups"`
	// the distance between limit and threshold
	MarginMb int `json:"margin_mb"`
	// when to handle event after receiving event
	Duration times.Duration `json:"duration"`
}

type MetricKind string

type MetricsCollectConfig struct {
	Node       MetricsNodeConfig      `json:"node"`
	Container  MetricsContainerConfig `json:"container"`
	Perf       MetricsPerfConfig      `json:"perf"`
	Rdt        MetricsRdtConfig       `json:"rdt"`
	Prometheus MetricsPrometheus      `json:"prometheus"`
}

type MetricsContainerConfig struct {
	Resources               []string       `json:"resources"`
	Cgroups                 []string       `json:"cgroups"`
	CollectInterval         times.Duration `json:"collect_interval"`
	MaxHousekeepingInterval times.Duration `json:"max_housekeeping_interval"`
}

type MetricsNodeConfig struct {
	CollectInterval times.Duration `json:"collect_interval"`
	SystemProcesses []string       `json:"system_processes"`
	OfflineType     string         `json:"-"`
	Devices         `json:",inline"`
}

type MetricsPerfConfig struct {
	Disable         bool           `json:"disable"`
	CollectInterval times.Duration `json:"collect_interval"`
	CollectDuration times.Duration `json:"collect_duration"`
	IgnoredCgroups  []string       `json:"ignored_cgroups"`
}

type MetricsPrometheus struct {
	CollectInterval times.Duration `json:"collect_interval"`
	// if need to show these metrics with the prefix "caelus_"
	DisableShow bool              `json:"disable_show"`
	Items       []*PrometheusData `json:"items"`
}

type MetricsRdtConfig struct {
	Disable         bool           `json:"disable"`
	RdtCommand      string         `json:"rdt_command"`
	CollectInterval times.Duration `json:"collect_interval"`
	CollectDuration times.Duration `json:"collect_duration"`
	ExecuteInterval times.Duration `json:"execute_interval"`
}

type MetricsSource struct {
	CheckInterval times.Duration `json:"check_interval"`
	// MetricsCommand is a command to get job's current metrics value, it must return the format data, like:
	// Its output is {"code":0,"msg":"success","data":[{"job_name":"","metric_name":"","key1":xx,"key2":xx,...}]}
	MetricsCommand []string `json:"metrics_command"`
	// if need to run chroot when executing metrics command
	CmdNeedChroot *bool `json:"cmd_need_chroot"`
	// MetricsURL is a url to get the job's metrics value, it must return the format data, like:
	// Its output is <slo>,<metrics>.
	MetricsURL string `json:"metrics_url"`
}

type NodeResourceConfig struct {
	Disable        bool           `json:"disable"`
	UpdateInterval times.Duration `json:"update_interval"`
	OfflineType    string         `json:"-"`
	// DisableKillIfNormal does not kill pod when no resource in conflicting status
	DisableKillIfNormal         bool                   `json:"disable_kill_if_normal"`
	OnlyKillIfIncompressibleRes bool                   `json:"only_kill_if_incompressible_res"`
	YarnConfig                  YarnNodeResourceConfig `json:"yarn_config"`
	Silence                     SilenceConfig          `json:"silence"`
}

type NotifyConfig struct {
	MemoryCgroup *MemoryNotifyConfig `json:"memory_cgroup"`
}

type OfflineJobs struct {
	Metadata interface{}
	Request  v1.ResourceList
	Used     v1.ResourceList
	State    string
}

type OnlineConfig struct {
	Enable       bool              `json:"enable"`
	PidToCgroup  PidToCgroup       `json:"pid_to_cgroup"`
	Jobs         []OnlineJobConfig `json:"jobs"`
	CustomMetric CustomMetric      `json:"custom_metric"`
}

type OnlineJobConfig struct {
	Name string `json:"name"`
	// JobCommand is job's command expression
	Command string          `json:"command"`
	Metrics []OnlineMetrics `json:"metrics"`
}

type OnlineMetrics struct {
	Name   string        `json:"name"`
	Source MetricsSource `json:"source"`
}

type OverCommit struct {
	Enable            bool                  `json:"enable"`
	OverCommitPercent float64               `json:"over_commit_percent"`
	Periods           []TimeRangeOverCommit `json:"periods"`
}

type PathInfo struct {
	Path string
	Size *DiskQuotaSize
	//if we set share limit, SharedInfo containers project id name
	//if not, SharedInfo is nil
	SharedInfo *SharedInfo
}

type PidToCgroup struct {
	// PidCheckInterval could be zero
	PidCheckInterval    times.Duration `json:"pids_check_interval"`
	CgroupCheckInterval times.Duration `json:"cgroup_check_interval"`
	BatchNum            int            `json:"batch_num"`
}

type PredictConfig struct {
	Disable       bool           `json:"disable"`
	CheckInterval times.Duration `json:"check_interval"`
	// PredictType must in [local, localv2, vpa]
	PredictType       string   `json:"predict_type"`
	PredictServerAddr string   `json:"predict_server_addr"`
	ReserveResource   Resource `json:"reserve_resource"`
	// PrintInterval is the the time interval to print predict detailed log for debug
	PrintInterval times.Duration `json:"print_interval"`
	// LocalPredictConfig is the configuration for local predictor
	LocalPredictConfig `json:",inline"`
	// The type value of online predict metrics caelus_node_resource{type=""}
	// It's used by experiment predict
	PredictMetricsType string `json:"predict_metrics_type"`
}

type PrometheusData struct {
	Address      string      `json:"address"`
	Collect      []string    `json:"collect"`
	NoCollect    []string    `json:"no_collect"`
	CollectMap   sets.String `json:"-"`
	NoCollectMap sets.String `json:"-"`
}

type RangeResource struct {
	CPUMilli RangeState `json:"cpu_milli"`
	MemMB    RangeState `json:"mem_mb"`
}

type RangeState struct {
	// Minimum is the range quantity
	Min float64 `json:"min"`
	// Maximum is the maxisum range quantity
	Max float64 `json:"max"`
	// Ratio used to calculate change range quantity
	Ratio float64 `json:"ratio"`
}

type RemoteAlarm struct {
	RemoteWebhook string `json:"remoteWebhook"`
	WeWorkWebhook string `json:"weWorkWebhook"`
}

type Resource struct {
	CpuMilli      *float64 `json:"cpu_milli"`
	MemMB         *float64 `json:"mem_mb"`
	CpuPercentStr string   `json:"cpu_percent"`
	CpuPercent    *float64 `json:"-"`
	MemPercentStr string   `json:"mem_percent"`
	MemPercent    *float64 `json:"-"`
}

type ResourceIsolateConfig struct {
	Disable         bool            `json:"disable"`
	ResourceDisable map[string]bool `json:"resource_disable"`
	UpdatePeriod    times.Duration  `json:"update_period"`
	// disks need to set io weight
	DiskNames []string `json:"-"`
	// eni iface for eni network pods
	EniIface string `json:"-"`
	// normal iface for host network and global route network pods
	Iface              string            `json:"-"`
	CpuConfig          CpuIsolateConfig  `json:"cpu_config"`
	OnlineType         string            `json:"-"`
	OfflineType        string            `json:"-"`
	ExternalComponents []ComponentConfig `json:"external_components"`
}

type ResourceUpdateEvent struct {
	ConflictRes []string
	Reason      string
}

type RoundOffResource struct {
	CPUMilli float64 `json:"cpu_milli"`
	MemMB    float64 `json:"mem_mb"`
}

type RuleCheck struct {
	ContainerRules []*RuleCheckConfig `json:"container_rules"`
	NodeRules      []*RuleCheckConfig `json:"node_rules"`
	AppRules       []*RuleCheckConfig `json:"app_rules"`
}

type RuleCheckConfig struct {
	Name    string   `json:"name"`
	Metrics []string `json:"metrics"`
	// CheckInterval describes the interval to trigger detection
	CheckInterval times.Duration `json:"check_interval"`
	// HandleInterval describes the interval to handle conflicts after detecting abnormal result
	HandleInterval times.Duration `json:"handle_interval"`
	// RecoverInterval describes the interval to recover conflicts after detecting normal result
	RecoverInterval times.Duration        `json:"recover_interval"`
	Rules           []*DetectActionConfig `json:"rules"`
	RecoverRules    []*DetectActionConfig `json:"recover_rules"`
}

type SharedInfo struct {
	PodName string
}

type SilenceConfig struct {
	// [0:00:00, 5:00:00]
	Periods [][2]times.SecondsInDay `json:"periods"`
	// disable schedule before silence
	AheadOfUnSchedule times.Duration `json:"ahead_of_unSchedule"`
}

type TaskTypeConfig struct {
	OnlineType  string `json:"online_type"`
	OfflineType string `json:"offline_type"`
}

type TimeRangeOverCommit struct {
	Range             [2]times.SecondsInDay `json:"range"`
	OverCommitPercent float64               `json:"over_commit_percent"`
}

type VolumeType string

type YarnDisksConfig struct {
	// RatioToCore translate disk space to core numbers
	RatioToCore      int64 `json:"ratio_to_core"`
	MultiDiskDisable bool  `json:"multi_disk_disable"`
	// DiskMinCapacityGb drop disks with little disk space
	DiskMinCapacityGb int64          `json:"disk_min_capacity_gb"`
	SpaceCheckEnabled bool           `json:"space_check_enabled"`
	SpaceCheckPeriod  times.Duration `json:"space_check_period"`
	// SpaceCheckReservedGb is used for checking disk space, it will start cleaning space if free disk space is less
	// than SpaceCheckReservedGb
	SpaceCheckReservedGb      int64   `json:"space_check_reserved_gb"`
	SpaceCheckReservedPercent float64 `json:"space_check_reserved_percent"`
	SpaceCleanDisable         bool    `json:"space_clean_disable"`
	// SpaceCleanJustData is enabled, it will just restart nodemanager pod to release /data space, and
	// do not care other disk partitions
	SpaceCleanJustData bool `json:"space_clean_just_data"`
	// OfflineExitedCleanDelay is used to clean nodemanager local or log path when offline pod exited for long time
	OfflineExitedCleanDelay times.Duration `json:"offline_exited_clean_delay"`
}

type YarnNodeResourceConfig struct {
	// CapacityIncInterval is used to make nodemanager capacity increase not very frequently
	CapacityIncInterval times.Duration    `json:"capacity_inc_interval"`
	NMServer            string            `json:"nm_server"`
	NMReserve           Resource          `json:"nm_reserve"`
	ResourceRoundOff    RoundOffResource  `json:"resource_roundoff"`
	ResourceRange       RangeResource     `json:"resource_range"`
	ScheduleServerPort  string            `json:"schedule_server_port"`
	PortAutoDetect      bool              `json:"port_auto_detect"`
	Properties          map[string]string `json:"properties"`
	Disks               YarnDisksConfig   `json:"disks"`
	ShimServer          string            `json:"shim_server"`
	CpuOverCommit       OverCommit        `json:"cpu_over_commit"`
}