Skip to content

Conversation

@toddpalino
Copy link

Evaluating partitions that lack complete offset data can lead to false positives. One particular case we saw that if there is only 1 offset for the partition, it will get incorrectly marked as stopped. There are other conceivable cases where you're going to get a bad evaluation until you have a baseline amount of data to work with. In order to deal with this, we can add a minimum threshold for completeness for evaluating partitions: any partition which has a completeness below this value will be marked as OK without evaluating it.

In order to make sure that you have at least 2 out of 10 offsets stored, you can set the minimum-completeness value to 0.20. The default value will be 0.0 (which is the current behavior). The configuration is set on the caching evaluator, so it goes under the config heading for that. You would need to define a config section like the following:

[evaluator.default]
class-name="caching"
minimum-complete=0.20

@coveralls
Copy link

coveralls commented May 2, 2018

Coverage Status

Coverage increased (+0.03%) to 74.002% when pulling 6efa004 on toddpalino:add-minimum-completeness into dc4cb84 on linkedin:master.

@toddpalino toddpalino merged commit 7c0b8b1 into linkedin:master May 2, 2018
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants