[ https://issues.apache.org/jira/browse/GERONIMO-3541?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12536881 ]
Erik B. Craig commented on GERONIMO-3541: ----------------------------------------- Viet - I might suggest a different approach to saving the information in the database. Currently you are using 1 row for every mbean/data element combination... with a default configuration on tomcat that is 46 rows - or roughly 8KB of data per snapshot with how your table is configured... with a default configuration snapshot interval of 5 minutes, that will result in significant memory usage rather quickly. I might suggest an alternative approach, breaking apart the mbeans into a separate table so that the raw text of the mbean path isn't re-saved thousands of times, as it it particularly redundant. I would also suggest breaking away the data element names into this table, perhaps in the same row as the mbean names delimited by a comma or a colon... I have a few other ideas which will result in significantly less storage and memory usage over time that we can discuss in more detail if you wish. > monitoring plugin: collecting agent needs to use a database to store the > snapshots instead of xml > ------------------------------------------------------------------------------------------------- > > Key: GERONIMO-3541 > URL: https://issues.apache.org/jira/browse/GERONIMO-3541 > Project: Geronimo > Issue Type: Improvement > Security Level: public(Regular issues) > Components: monitoring > Affects Versions: 2.1 > Environment: windows > Reporter: Viet Hung Nguyen > Attachments: geronimo-3541.patch > > > The DB needs to be populated with snapshot statistics. Also, the functions > used to query snapshots needs to be migrated from reading from XML files to a > DB. -- This message is automatically generated by JIRA. - You can reply to this email to add a comment to the issue online.