[CASSANDRA-15661] Improve logging by using more appropriate levels

• Type: Improvement
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 4.0-alpha4, 4.0
• Component/s: Observability/Logging
• Labels: pull-request-available
• Change Category: Operability
• Complexity: Low Hanging Fruit
• Platform: All

Source Control Link

https://github.com/apache/cassandra/commit/dd9a6e0a8a691bdee357eb6559f1b2b92b076925

Test and Documentation Plan

CircleCI

• Assignee: Jon Haddad
• Reporter: Jon Haddad

Authors

Jon Haddad

Reviewers

Alexander Dejanovski, Jon Haddad

• Created: 24/Mar/20 20:35
• Updated: 15/May/20 08:54
• Resolved: 01/Apr/20 15:19

Description

There are a number of log statements using logging levels that are a bit too conservative. For example:

• Flushing memtables is currently at debug. This is a relatively rare event that is important enough to be INFO
• When compaction finishes we log the progress at debug
• Different steps in incremental repair are logged as debug, should be INFO
• when reaching connection limits in ConnectionLimitHandler.java we log at warn rather than error. Since this is a client disconnect it’s more than a warning, we’re taking action and disconnecting.

Issue Links

• links to: GitHub Pull Request #488