On socket failure, re-initialize the address info.
Along with #468, this triggers a new DNS lookup.
To prevent immediate retries that quickly exhaust the retry count,
add a retry_delay_ms to FluentdExporterOptions, defaulting to 1000ms.
We found this useful when deploying in a Kubernetes environment,
where the targeted service endpoint may not yet be up when a logging client starts up.
On socket failure, re-initialize the address info.
Along with #468, this triggers a new DNS lookup.
To prevent immediate retries that quickly exhaust the retry count, add a retry_delay_ms to FluentdExporterOptions, defaulting to 1000ms.
We found this useful when deploying in a Kubernetes environment, where the targeted service endpoint may not yet be up when a logging client starts up.