Basic settings and options

Retry

If the response to a gRPC call corresponds to a configured error, the call is retried. The retry will be executed according to the configured Retry settings. The default retry responses and settings for each API are as follows.

Retry error status code

For more information, please refer to GRPC Github.

Item	Description
CANCELLED	If the call is canceled (usually canceled by the calling party)
UNKNOWN	This error can be returned for unknown errors, for example, if a status value received from another address is not defined for that address - Also for APIs that do not provide information about response errors
DEADLINE_EXCEEDED	The deadline expired before the call completed
NOT_FOUND	Some requested entity (e.g. a file or path) could not be found
RESOURCE_EXHAUSTED	Some resource is exhausted, or the user quota is insufficient, or the file system as a whole is out of space
ABORTED	Usually when a sequencer check fails, or the operation is aborted due to a concurrency issue, such as in the middle of a transaction
INTERNAL	Internal error
UNAVAILABLE	The service is currently unavailable

Retry settings

Item	Description
Jittered	Whether to randomly set the delay time - Most jitters are set to true
InitiAlertryDelay	Delays to set before first retry - Use a value adjusted based on RetryDelayMultiplier
RetryDelayMultiplier	Next call retry delay = previous call retry delay * RetryDelayMultiplier
MaxRetryDelay	Sets a limit on retry delays - RetryMultiplier cannot set retry delays exceeding this value
InitialRpcTimeout	Set a timeout for the first RPC - Use a value adjusted based on RpcTimeoutMultiplier
RpcTimeoutMultiplier	Next call timeout = previous call timeout * RpcTimeoutMultiplier
MaxRpcTimeout	Sets the RPC timeout - RpcTimeoutMultiplier cannot set the RPC timeout exceeding this value
TotalTimeout	Ultimately controls how long we should keep trying remote calls before giving up on the logic entirely - The higher this value, the more retries we will execute

Publisher retry settings

API	Retry code	Description	Retry setting
create topic	14	UNAVAILABLE	• Initial : 100ms • Multiplier : 1.3 • Max : 60s
get topic	2	UNKNOWN
	10	ABORTED
	14	UNAVAILABLE
list topics	2	UNKNOWN
	10	ABORTED
	14	UNAVAILABLE
list topic subscriptions	2	UNKNOWN
	10	ABORTED
	14	UNAVAILABLE
update topic	14	UNAVAILABLE
delete topic	14	UNAVAILABLE
publish	1	CANCELLED
	2	UNKNOWN
	4	DEADLINE_EXCEEDED
	8	RESOURCE_EXHAUSTED
	10	ABORTED
	13	INTERNAL
	14	UNAVAILABLE

Subscriber retry settings

API	Retry code	Description	Retry setting
create subscription	14	UNAVAILABLE	• Initial: 100ms • Multiplier: 1.3 • Max: 60s
get subscription	14	UNAVAILABLE
list subscriptions	14	UNAVAILABLE
seek	14	UNAVAILABLE
update subscription	14	UNAVAILABLE
delete subscription	14	UNAVAILABLE
modify ack deadline	14	UNAVAILABLE
acknowledge	14	UNAVAILABLE
pull	14	UNAVAILABLE
streaming pull	4	DEADLINE_EXCEEDED
	8	RESOURCE_EXHAUSTED
	13	INTERNAL
	14	UNAVAILABLE

Batch processing

When making an API call, messages can be bundled and sent in batches.

Limit Exceeded Behavior

Action to take when the flow control limit is exceeded

• SignalError: Returns an error.
• Block: Waits until the request can be made without exceeding the limit. Useful in batch processing where the latency of individual requests is not important.
• Ignore: Does not use flow control.

API	Batching Settings	Element Count Threshold	Request Bytes Threshold	Delay Threshold	Limit Exceeded Behavior
publish	100	1000 (=1KB)	10ms	SignalError

Logging

You can leave logs for requests and responses (success/failure) in Publish and StreamingPull as statistics.

• If you set a time interval, the number of successes/failures of publish and pull will be output for each interval.
• If you want to check the output, you must set it to info level or higher.

The following values ​​can be set for Logger by default.

Item	Description
Level	Log level - Default: Panic
Output	Log output - Default: os.Stdout
StatLogInterval	Time interval to leave logs

Logging example

conf := &ClientConfig{
    LogOptions: &LogOptions{
        Level: "info",
        Output: os.Stdout,
        StatLogInterval: 1 * time.Second
    }
}
client, err := NewClientWithConfig(ctx, domainID, projectID, conf, opts...)