Although ICEpush supports automatic retries for failed heartbeat connections, we don’t do the same for failed push update retrieval attempts. Instead, we treat these the same as user-initiated requests that will immediately fail due to a network error.
It would be friendlier if IcePush would handle failed push update retrievals by automatically retrying in a manner similar to that used for the existing heartbeat connections.
The existing org.icepush.networkErrorRetryTimeouts configuration could be used for both heartbeat connection failures and push update retrieval failures.
This change is especially important for push update retrieval requests which happen behind the scenes unbeknownst to the user. If I was a user, I might be confused by an asynchronous NCI popup that blocks further operations on my window due to a failed background ajax update which, if retried, would likely succeed.