telemetry/exporter/exporterhelper/README.md

Exporter Helper

This package provides reusable implementations of common capabilities for exporters. Currently, this includes queuing, batching, timeouts, and retries.

Configuration

The following configuration options can be modified:

Retry on Failure

• retry_on_failure
  • enabled (default = true)
  • initial_interval (default = 5s): Time to wait after the first failure before retrying; ignored if enabled is false
  • max_interval (default = 30s): Is the upper bound on backoff; ignored if enabled is false
  • max_elapsed_time (default = 300s): Is the maximum amount of time spent trying to send a batch; ignored if enabled is false. If set to 0, the retries are never stopped.
  • multiplier (default = 1.5): Factor by which the retry interval is multiplied on each attempt; ignored if enabled is false

Sending Queue

• sending_queue
  • enabled (default = true)
  • num_consumers (default = 10): Number of consumers that dequeue batches; ignored if enabled is false
  • wait_for_result (default = false): determines if incoming requests are blocked until the request is processed or not.
  • block_on_overflow (default = false): If true, blocks the request until the queue has space otherwise rejects the data immediately; ignored if enabled is false
  • sizer (default = requests): How the queue and batching is measured. Available options:
    • requests: number of incoming batches of metrics, logs, traces (the most performant option);
    • items: number of the smallest parts of each signal (spans, metric data points, log records);
    • bytes: the size of serialized data in bytes (the least performant option).
  • queue_size (default = 1000): Maximum size the queue can accept. Measured in units defined by sizer
  • batch: see below.

Failure behavior: If data cannot be added to the sending queue, it is typically dropped. This occurs when the queue has reached its configured capacity or, for persistent queues, when the underlying storage cannot accept additional data (for example, due to insufficient disk space or I/O errors).

When block_on_overflow is enabled, the caller may instead wait until space becomes available, and the request may still be enqueued if capacity frees up before the timeout.

If data is rejected before entering the queue, it does not reach the exporter retry logic. Such enqueue failures are reported by the otelcol_exporter_enqueue_failed_* metrics.

Sending queue batch settings

Batch settings are available in the sending queue. Batching is disabled, by default. To enable default batch settings, use batch: {}. When batch is defined, the settings are:

• flush_timeout (default = 200 ms): time after which a batch will be sent regardless of its size. Must be a non-zero value;
• min_size (default = 8192): the minimum size of a batch; should be less than or equal to the sending_queue::queue_size if sending_queue::batch::sizer matches sending_queue::sizer.
• max_size (default = 0): the maximum size of a batch, enables batch splitting. The maximum size of a batch should be greater than or equal to the minimum size of a batch. If set to zero, there is no maximum size;
• sizer: see below.
• partition: see below.

The batch::sizer field is given special treatment because the queue itself also defines a sizer. This field supports using different size limits for the queue and batch-related logic.

If the batch::sizer field is not set, it takes its value from the parent structure.

If sending_queue::sizer is not set, batch::sizer defaults to items.

Available batch::sizer options:

• items: number of the smallest parts of each signal (spans, metric data points, log records);
• bytes: the size of serialized data in bytes (the least performant option).

The batch::partition configuration defines the partitioning of the batches.

Available batch::partition options:

• metadata_keys: a list of client.Metadata keys used to partition data into separate batches. When empty, a single batcher instance is used. When set, one batcher will be used per distinct combination of values for the listed metadata keys. Empty value and unset metadata are treated as distinct cases. Entries are case-insensitive. Duplicated entries will trigger a validation error. Default is empty.

Timeout

• timeout (default = 5s): Time to wait per individual attempt to send data to a backend

The initial_interval, max_interval, max_elapsed_time, and timeout options accept duration strings, valid time units are "ns", "us" (or "µs"), "ms", "s", "m", "h".

Persistent Queue

To use the persistent queue, the following setting needs to be set:

• sending_queue
  • storage (default = none): When set, enables persistence and uses the component specified as a storage extension for the persistent queue. There is no in-memory queue when set.

The maximum number of batches stored to disk can be controlled using sending_queue.queue_size parameter (which, similarly as for in-memory buffering, defaults to 1000 batches).

When persistent queue is enabled, the batches are being buffered using the provided storage extension - filestorage is a popular and safe choice. If the collector instance is killed while having some items in the persistent queue, on restart the items will be picked and the exporting is continued.

Context Propagation: Request context (including client metadata and span context) is preserved when using persistent queues. However, context set by Auth extensions is not propagated through the persistent queue. Auth extension context is ignored when data is persisted to disk, which means authentication/authorization information will not be available when the persisted data is processed.

                                                              ┌─Consumer #1─┐
                                                              │    ┌───┐    │
                              ──────Deleted──────        ┌───►│    │ 1 │    ├───► Success
        Waiting in channel    x           x     x        │    │    └───┘    │
        for consumer ───┐     x           x     x        │    │             │
                        │     x           x     x        │    └─────────────┘
                        ▼     x           x     x        │
┌─────────────────────────────────────────x─────x───┐    │    ┌─Consumer #2─┐
│                             x           x     x   │    │    │    ┌───┐    │
│     ┌───┐     ┌───┐ ┌───┐ ┌─x─┐ ┌───┐ ┌─x─┐ ┌─x─┐ │    │    │    │ 2 │    ├───► Permanent -> X
│ n+1 │ n │ ... │ 6 │ │ 5 │ │ 4 │ │ 3 │ │ 2 │ │ 1 │ ├────┼───►│    └───┘    │      failure
│     └───┘     └───┘ └───┘ └───┘ └───┘ └───┘ └───┘ │    │    │             │
│                                                   │    │    └─────────────┘
└───────────────────────────────────────────────────┘    │
   ▲              ▲     ▲           ▲                    │    ┌─Consumer #3─┐
   │              │     │           │                    │    │    ┌───┐    │
   │              │     │           │                    │    │    │ 3 │    ├───► (in progress)
 write          read    └─────┬─────┘                    ├───►│    └───┘    │
 index          index         │                          │    │             │
                              │                          │    └─────────────┘
                              │                          │
                          currently                      │    ┌─Consumer #4─┐
                          dispatched                     │    │    ┌───┐    │     Temporary
                                                         └───►│    │ 4 │    ├───►  failure
                                                              │    └───┘    │         │
                                                              │             │         │
                                                              └─────────────┘         │
                                                                     ▲                │
                                                                     └── Retry ───────┤
                                                                                      │
                                                                                      │
                                                   X  ◄────── Retry limit exceeded ───┘

Example:

receivers:
  otlp:
    protocols:
      grpc:
exporters:
  otlp_grpc:
    endpoint: <ENDPOINT>
    sending_queue:
      storage: file_storage/otc
extensions:
  file_storage/otc:
    directory: /var/lib/storage/otc
    timeout: 10s
service:
  extensions: [file_storage]
  pipelines:
    metrics:
      receivers: [otlp]
      exporters: [otlp]
    logs:
      receivers: [otlp]
      exporters: [otlp]
    traces:
      receivers: [otlp]
      exporters: [otlp]