View Source NervesHubLinkCommon.Downloader.RetryConfig (nerves_hub_link_common v0.4.2)
Configuration structure for how the Downloader server will handle disconnects, errors, timeouts etc
Link to this section Summary
Types
time in milliseconds between chunks of data received
that once elapsed will trigger a retry. This event counts
towards the max_disconnects counter
maximum number of disconnects. After this limit is reached the download will be stopped and will no longer be retried
maximum time in milliseconds that a download can exist for. after this amount of time has elapsed, the download is canceled and the download process will crash
time in milliseconds to wait before attempting to retry a download
worst case download speed specified in bytes per second. This is
used to calculate the "worst case" download timeout. it is meant to
fail faster than waiting for max_timeout to elapse
Link to this section Types
@type idle_timeout() :: non_neg_integer()
time in milliseconds between chunks of data received
that once elapsed will trigger a retry. This event counts
towards the max_disconnects counter
@type max_disconnects() :: non_neg_integer()
maximum number of disconnects. After this limit is reached the download will be stopped and will no longer be retried
@type max_timeout() :: non_neg_integer()
maximum time in milliseconds that a download can exist for. after this amount of time has elapsed, the download is canceled and the download process will crash
@type t() :: %NervesHubLinkCommon.Downloader.RetryConfig{ idle_timeout: idle_timeout(), max_disconnects: max_disconnects(), max_timeout: max_timeout(), time_between_retries: time_between_retries(), worst_case_download_speed: worst_case_download_speed() }
@type time_between_retries() :: non_neg_integer()
time in milliseconds to wait before attempting to retry a download
@type worst_case_download_speed() :: non_neg_integer()
worst case download speed specified in bytes per second. This is
used to calculate the "worst case" download timeout. it is meant to
fail faster than waiting for max_timeout to elapse