[HBASE-17722] Metrics subsystem stop/start messages add a lot of useless bulk to operational logging

• Type: Bug
• Status: Resolved
• Priority: Trivial
• Resolution: Fixed
• Affects Version/s: 1.3.0, 1.2.4
• Fix Version/s: 1.4.0, 1.3.1, 1.1.10, 1.2.6, 2.0.0
• Component/s: metrics
• Assignee: Andrew Kyle Purtell
• Reporter: Andrew Kyle Purtell
• Created: 03/Mar/17 01:56
• Updated: 29/May/17 05:05
• Resolved: 03/Mar/17 20:41

Description

Metrics subsystem stop/start messages add a lot of useless bulk to operational logging. Say you are collecting logs from a fleet of thousands of servers and want to have them around for ~month or longer. It adds up.

I think these should at least be at DEBUG level and ideally at TRACE. They don't offer much utility. Unfortunately they are Hadoop classes so we can tweak log4j.properties defaults instead. We do this in test resources but not in what we ship in conf/ .

<time> INFO  [<thread>] impl.MetricsSystemImpl: HBase metrics system started

<time> INFO  [<thread>] impl.MetricsSystemImpl: Stopping HBase metrics system...

<time> INFO  [<thread>] impl.MetricsSystemImpl: HBase metrics system stopped.

<time> INFO  [<thread>] impl.MetricsConfig: loaded properties from hadoop-metrics2-hbase.properties

<time> INFO  [<thread>] impl.MetricsSystemImpl: Scheduled snapshot period at 10 second(s).