[
https://issues.apache.org/jira/browse/STREAMS-87?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14002441#comment-14002441
]
ASF GitHub Bot commented on STREAMS-87:
---------------------------------------
Github user smashew commented on a diff in the pull request:
https://github.com/apache/incubator-streams/pull/19#discussion_r12817778
--- Diff:
streams-contrib/streams-provider-twitter/src/main/java/org/apache/streams/twitter/provider/TwitterTimelineProvider.java
---
@@ -116,10 +116,8 @@ protected void captureTimeline(long currentId) {
int keepTrying = 0;
// keep trying to load, give it 5 attempts.
- //while (keepTrying < 10)
- while (keepTrying < 1)
+ while (keepTrying < 5)
--- End diff --
I mean, it could be user specified, but it is typically better to assume
that a 'service' could have an issue on first attempt. There are a lot of
reasons, legitimately that it would fail, including the service being
unavailable, too many questions being asked by the user, etc. 5 is a pretty
good default to let the backoff do it's job and wait for the service to become
available again.
Personally, I'm not a fan of 'magic numbers' either, but, something > 1 is
really necessary to be fault tolerant...
> TwitterTimelineProvider needs to allow for backoff
> --------------------------------------------------
>
> Key: STREAMS-87
> URL: https://issues.apache.org/jira/browse/STREAMS-87
> Project: Streams
> Issue Type: Bug
> Reporter: Robert Douglas
> Original Estimate: 1h
> Remaining Estimate: 1h
>
> At the moment, the captureTwitterTimeline function in TwitterTimelineProvider
> only tries one attempt at fetching a user's timeline. In the event of an
> exception being thrown, we simply drop the data we were trying to collect. We
> should instead attempt 5 times in order to allow enough time for Twitter's
> backoff
--
This message was sent by Atlassian JIRA
(v6.2#6252)
