memorylimiter

README

Memory Limiter Processor

Supported pipeline types: metrics, traces

The memory limiter processor is used to prevent out of memory situations on the collector. Given that the amount and type of data a collector processes is environment specific and resource utilization of the collector is also dependent on the configured processors, it is important to put checks in place regarding memory usage. The memory_limiter processor offers the follow safeguards:

• Ability to define an interval when memory usage will be checked and if memory usage exceeds a defined limit will trigger GC to reduce memory consumption.
• Ability to define an interval when memory usage will be compared against the previous interval's value and if the delta exceeds a defined limit will trigger GC to reduce memory consumption.

In addition, there is a command line option (mem-ballast-size-mib) which can be used to define a ballast, which allocates memory and provides stability to the heap. If defined, the ballast increases the base size of the heap so that GC triggers are delayed and the number of GC cycles over time is reduced. While the ballast is configured via the command line, today the same value configured on the command line must also be defined in the memory_limiter processor.

Note that while these configuration options can help mitigate out of memory situations, they are not a replacement for properly sizing and configuring the collector. For example, if the limit or spike thresholds are crossed, the collector will return errors to all receive operations until enough memory is freed. This may result in dropped data.

It is highly recommended to configure the ballast command line option as well as the memory_limiter processor on every collector. The ballast should be configured to be 1/3 to 1/2 of the memory allocated to the collector. The memory_limiter processor should be the first processor defined in the pipeline (immediately after the receivers). This is to ensure that backpressure can be sent to applicable receivers and minimize the likelihood of dropped data when the memory_limiter gets triggered.

Please refer to config.go for the config spec.

The following configuration options must be changed:

• check_interval (default = 0s): Time between measurements of memory usage. Values below 1 second are not recommended since it can result in unnecessary CPU consumption.
• limit_mib (default = 0): Maximum amount of memory, in MiB, targeted to be allocated by the process heap. Note that typically the total memory usage of process will be about 50MiB higher than this value.
• spike_limit_mib (default = 0): Maximum spike expected between the measurements of memory usage. The value must be less than limit_mib.

The following configuration options can also be modified:

• ballast_size_mib (default = 0): Must match the mem-ballast-size-mib command line option.

Examples:

processors:
  memory_limiter:
    ballast_size_mib: 2000
    check_interval: 5s
    limit_mib: 4000
    spike_limit_mib: 500

Refer to config.yaml for detailed examples on using the processor.

Documentation

Overview

Package memorylimiter provides a processor for OpenTelemetry Service pipeline that drops data on the pipeline according to the current state of memory usage.

Index

• type Config
• type DualTypeProcessor
• type Factory
  • func (f *Factory) CreateDefaultConfig() configmodels.Processor
  • func (f *Factory) CreateMetricsProcessor(_ context.Context, params component.ProcessorCreateParams, ...) (component.MetricsProcessor, error)
  • func (f *Factory) CreateTraceProcessor(_ context.Context, params component.ProcessorCreateParams, ...) (component.TraceProcessor, error)
  • func (f *Factory) Type() configmodels.Type

Types

type Config

type Config struct {
	configmodels.ProcessorSettings `mapstructure:",squash"`

	// CheckInterval is the time between measurements of memory usage for the
	// purposes of avoiding going over the limits. Defaults to zero, so no
	// checks will be performed.
	CheckInterval time.Duration `mapstructure:"check_interval"`

	// MemoryLimitMiB is the maximum amount of memory, in MiB, targeted to be
	// allocated by the process.
	MemoryLimitMiB uint32 `mapstructure:"limit_mib"`

	// MemorySpikeLimitMiB is the maximum, in MiB, spike expected between the
	// measurements of memory usage.
	MemorySpikeLimitMiB uint32 `mapstructure:"spike_limit_mib"`

	// BallastSizeMiB is the size, in MiB, of the ballast size being used by the
	// process.
	BallastSizeMiB uint32 `mapstructure:"ballast_size_mib"`
}

Config defines configuration for memory memoryLimiter processor.

type DualTypeProcessor

type DualTypeProcessor interface {
	consumer.TraceConsumer
	consumer.MetricsConsumer
	component.Processor
}

type Factory

type Factory struct {
}

Factory is the factory for Attribute Key processor.

func (*Factory) CreateDefaultConfig

func (f *Factory) CreateDefaultConfig() configmodels.Processor

CreateDefaultConfig creates the default configuration for processor. Notice that the default configuration is expected to fail for this processor.

func (*Factory) CreateMetricsProcessor

func (f *Factory) CreateMetricsProcessor(
	_ context.Context,
	params component.ProcessorCreateParams,
	nextConsumer consumer.MetricsConsumer,
	cfg configmodels.Processor,
) (component.MetricsProcessor, error)

CreateMetricsProcessor creates a metrics processor based on this config.

func (*Factory) CreateTraceProcessor

func (f *Factory) CreateTraceProcessor(
	_ context.Context,
	params component.ProcessorCreateParams,
	nextConsumer consumer.TraceConsumer,
	cfg configmodels.Processor,
) (component.TraceProcessor, error)

CreateTraceProcessor creates a trace processor based on this config.

func (*Factory) Type

func (f *Factory) Type() configmodels.Type

Type gets the type of the config created by this factory.