[ https://issues.apache.org/jira/browse/BEAM-400?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15365541#comment-15365541 ]
Daniel Halperin edited comment on BEAM-400 at 7/7/16 3:31 AM: -------------------------------------------------------------- Runner names are not helpful here. I'm going to assume you mean: * {code}PubsubIO.Read{code} * {code}Bounded{code} mode, using {code}withMaxNumRecords{code} or {code}withMaxReadTime{code}, or both. This would seem to be a potential improvement that could be made to {code}PubsubReader{code}. was (Author: dhalp...@google.com): Runner names are not helpful here. I'm going to assume you mean: * {code}PubsubIO.Read{code} * {code}Bounded{code} mode, using {code}withMaxNumRecords{code} or {code}withMaxReadTime{code}, or both. This would seem to be a potential improvement that could be made to {code}PubsubReader{code}. > Direct/InProcess runner should use Pubsub publish time timestamp if > timestampLabel is not set > --------------------------------------------------------------------------------------------- > > Key: BEAM-400 > URL: https://issues.apache.org/jira/browse/BEAM-400 > Project: Beam > Issue Type: Bug > Components: sdk-java-gcp > Reporter: Daniel Mills > Priority: Minor > > When using system-provided timestamps for PubsubIO, the timestamp should be > the time at which the message was published to Pubsub, not when it was > received by the pipeline. -- This message was sent by Atlassian JIRA (v6.3.4#6332)