[
https://issues.apache.org/jira/browse/CAMEL-4876?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13726751#comment-13726751
]
Preben Asmussen edited comment on CAMEL-4876 at 8/1/13 7:54 PM:
----------------------------------------------------------------
[~davsclaus]
Another way this could be usefull ->
If repeated errors occur, the poll could use the backoffMultiplier feature to
auto delay the next poll.
We could introduce options to control this e.g. autoBackOffOnError=true/false,
numberOfErrorsBeforeBackoff in combination with the other new backOff options.
When combining this with consumer.bridgeErrorHandler you would have a nice way
for consumer endpoints to deal with failing infrastructure.
I would even go so far to say that autoBackOff should be enabled by default on
3.0
was (Author: preben):
[~davsclaus]
Another way this could be use full ->
If repeated errors occur, the poll could use the backoffMultiplier feature to
auto delay the next poll.
We could introduce options to control this e.g. autoBackOffOnError=true/false,
numberOfErrorsBeforeBackoff in combination with the other new backOff options.
When combining this with consumer.bridgeErrorHandler you would have a nice way
for consumer endpoints to deal with failing infrastructure.
I would even go so far to say that autoBackOff should be enabled by default on
3.0
> Add support for a "back-off multiplier" capability to the
> ScheduledPollConsumer
> -------------------------------------------------------------------------------
>
> Key: CAMEL-4876
> URL: https://issues.apache.org/jira/browse/CAMEL-4876
> Project: Camel
> Issue Type: New Feature
> Reporter: Ashwin Karpe
> Fix For: 2.12.0
>
>
> Usually files or tables are only updated once a day or even once a week in a
> batch like fashion. When this happens its of course important to process as
> fast as possible (using the default 500 ms delay), but most of the time when
> there is no activity, polling every 500 ms. is not necessary and takes system
> resources when running many polling routes on the same box.
> I was thinking that the ScheduledPollConsumer could be more dynamic by
> introducing a new option eg. backoffMultiplier, that resets the scheduler to
> maxDelay if a poll results in no exchange (maybe after x polls with no
> results).
> The same goes if a poll results in an exchange, and the delay currently is at
> backoffMultiplier the scheduler is reset to the original delay thereby
> polling more agresive again.
> Original Camel User Forum request :
> http://camel.465427.n5.nabble.com/DISCUSS-Dynamic-ScheduledPollConsumer-td5129231.html
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
