You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@phoenix.apache.org by "Daniel Wong (Jira)" <ji...@apache.org> on 2020/10/03 08:14:00 UTC

[jira] [Created] (PHOENIX-6174) Phoenix Metrics Initial Unification

Daniel Wong created PHOENIX-6174:
------------------------------------

             Summary: Phoenix Metrics Initial Unification
                 Key: PHOENIX-6174
                 URL: https://issues.apache.org/jira/browse/PHOENIX-6174
             Project: Phoenix
          Issue Type: Improvement
    Affects Versions: 5.1.0, 4.x
            Reporter: Daniel Wong


I see this as an incremental improvement toward a larger more powerful metrics framework in phoenix-server, phoenix-client, and phoenix-connectors.  Today we describe our available metrics as https://phoenix.apache.org/metrics.html

Long term a tag based multi-dimensional data model (similar to prometheus/open tsdb) may be superior but for now this targets simple generalization of the existing metrics in phoenix.  Today much of the metrics are built on hbase/hadoop offering.

 

Some initial discussion/brainstorming:


Metric Types:

  Phoenix uses/provides sufficient verbosity of metrics, but does not define the metric types cleanly.  Typical counter, gauge, histogram

 

Monitoring Models:

  Events (Event Metrics), metrics with no summarization or aggregation.  

     Example of this will be request metrics.  On this request you used 5 scanners.

  Metrics (Summary/Aggregated Metrics), 

    Example of this might be total queries

Reporting:
  Today JMX is one of the main approaches used in phoenix for metrics.  For events most of them are retrievable from PhoenixRuntime.  Long term these should be pluggable.

 

Storage:

  End user based on reporting is envisioned.

  May consider storing in HBase/Phoenix as an option?  The query log could store events with that query log entry.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)