Add retention job summary events #3683
Conversation
| DatasetCleaner.this.finishCleanSignal.get().countDown(); | ||
| LOG.info("Successfully cleaned: " + dataset.datasetURN()); | ||
| DatasetRetentionSummary summary = new DatasetRetentionSummary(dataset.datasetURN(), datasetsDeleted, true); |
There was a problem hiding this comment.
is it certain we'd only ever overwrite the summary, rather than potentially needing to add this new count to whatever accumulating sum was already tracked for the same dataset URN?
| * Cleans the {@link CleanableDataset}. | ||
| * Returns the number of versions marked for deletion | ||
| * {@link org.apache.gobblin.data.management.retention.policy.RetentionPolicy} and delete files and directories that need deleting. |
There was a problem hiding this comment.
looks like new line 33 cut in mid-sentence to the prior javadoc.
also, prefer a @return tag
| private final int versionsSelectedForDeletion; | ||
| private final boolean successfullyDeleted; |
There was a problem hiding this comment.
in the impl, it appeared that versionsSelectedForDeletion is truly the total number successfully deleted. e.g. is there any way that the number selected might not equal a non-zero number successfully deleted?
if not, I suggest renaming to versionsDeleted and reversing the boolean to hadExecutionFailure. the latter of course discerns between 0 found to delete and >0 found, but none actually successful. calling out a tad more explicitly in javadoc seems worthwhile
| @@ -34,6 +35,12 @@ static class CleanFailed { | |||
| */ | |||
| static final String FAILURE_CONTEXT_METADATA_KEY = "failureContext"; | |||
| } | |||
|
|
|||
| static class RetentionJobSummary { | |||
There was a problem hiding this comment.
nit: RetentionEvents.JobSummary would avoid redundant Retention
Dear Gobblin maintainers,
Please accept this PR. I understand that it will not be reviewed until I have checked off all the steps below!
JIRA
Description
Tests
Commits