Retry and Queueing

Sending Queue

A sending queue is a buffer that stores telemetry data temporarily before sending it to the destination. The sending queue ensures that telemetry data is not lost due to network connectivity issues or server outages and helps to minimize the number of network connections required for efficient transmission.

Parameter	Type	Default	Description
sending_queue_enabled	bool	true	Enable to buffer telemetry data temporarily before sending
sending_queue_num_consumers	int	10	The number of consumers that dequeue batches.
sending_queue_queue_size	int	5000	Maximum number of batches kept in memory before dropping.

Persistent Queue

In addition to the sending queue, the persistent queue may be enabled. When enabled, telemetry data is persisted to the disk, which provides data resiliency in cases where the collector restarts. The sending queue must be enabled to enable persistent queuing.

Parameter	Type	Default	Description
persistent_queue_enabled	bool	true	Enable to buffer telemetry data to disk instead of in memory.
persistent_queue_directory	string	$OIQ_OTEL_COLLECTOR_HOME/storage	The path to a directory where telemetry data will be buffered.

Retry on Failure

Retry on failure settings are used to determine whether the exporter should attempt to resend telemetry data that has failed to be transmitted to the destination endpoint. When this setting is enabled, the exporter will automatically retry failed transmissions at a configurable interval until the data is successfully transmitted. This helps to ensure that telemetry data is not lost due to temporary network connectivity issues or server outages.

Parameter	Type	Default	Description
retry_on_failure_enabled	bool	true	Attempt to resend telemetry data that has failed to be transmitted to the destination.
retry_on_failure_initial_interval	int	5	Time (in seconds) to wait after the first failure before retrying.
retry_on_failure_max_interval	int	30	The upper bound (in seconds) on backoff.
retry_on_failure_max_elapsed_time	int	300	The maximum amount of time (in seconds) spent trying to send a batch, used to avoid a never-ending retry loop. When set to 0, the retries are never stopped.