[ https://issues.apache.org/jira/browse/AMBARI-21145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16251852#comment-16251852 ]
Juanjo Marron commented on AMBARI-21145: ---------------------------------------- Thanks for the update! > Allow wildcard for log directory folder in the path component of Logfeeder > input > -------------------------------------------------------------------------------- > > Key: AMBARI-21145 > URL: https://issues.apache.org/jira/browse/AMBARI-21145 > Project: Ambari > Issue Type: Improvement > Components: logsearch > Affects Versions: 2.5.0 > Reporter: Keta Patel > Assignee: Olivér Szabó > Priority: Minor > Fix For: 2.6.1 > > > The wildcard processing in Logfeeder is carried out only for the last file > component in the log path. For a few services, where there are multiple log > folders and log files from each of the folders is desired to be parsed in > Logsearch UI, the lack of a wildcard processing for the previous portions > becomes an issue. Hence, introducing wildcard processing for the log > directory part of the "path" in the Logfeeder's input for the service will > help resolve this issue. > The idea is if we have a folder pattern, we will craete a thread group, and > there will be different threads for every matching input (now only for tail > file) > Features (new input block settings): > - {{path_update_interval_min}}: the period in minutes for checking new files > (default: 5, based on detach values, its possible that a new input wont be > monitored) > - {{detach_interval_min}}: the period in minutes for checking which files are > too old (default: 300) > - {{detach_time_min: the period}} in minutes when we flag a file is too old > (default: 2000) > - if a path contains a {{*}} before the last {{/}}, then logfeeder will know > that the inputs should handled in a thread group, not in a standalone thread -- This message was sent by Atlassian JIRA (v6.4.14#64029)