[jira] [Commented] (IOTDB-1761) Add metric framework for IoTDB

["xinzhongtianxia (Jira)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/IOTDB-1761?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17438709#comment-17438709 ] 

xinzhongtianxia commented on IOTDB-1761:
----------------------------------------

Glad to see this feature and it's helpful to improve an online-system's stability。
Recently, I prepare to build our own system to monitor a running IoTDB.
But after a looking at IoTDB's codes, I have some doubt, could any one make it a bit clear? Many many thanks.
1. I found there already are some implementations related to metrics, such as the metrics and cost.statistic package in server module and the monitor package both in server and cluster module.
Will they finally all be removed after this feature's complementation?
2. There also are some MBeans, Will they be removed too? After all, the new metric framework could export data to JMX。
IMO, It is better to have only one metric framework in IoTDB。

> Add metric framework for IoTDB
> ------------------------------
>
>                 Key: IOTDB-1761
>                 URL: https://issues.apache.org/jira/browse/IOTDB-1761
>             Project: Apache IoTDB
>          Issue Type: New Feature
>          Components: Core/Others
>            Reporter: Steve Yurong Su
>            Assignee: Steve Yurong Su
>            Priority: Major
>              Labels: pull-request-available
>   Original Estimate: 16h
>  Remaining Estimate: 16h
>
> The system has two characteristics:
>  # Define the interface of measurement collection system, including measurement collection, and measurement data exposure. You can customize your own implementation based on these interfaces.
>  ## There are five kinds of measurement collection interfaces
>  1. Counter: can only be incremented
>  2. Gauge: instantaneous value, can be set and obtained
>  3. Rate: throughput, support 1 minute, 5 minutes, 15 minutes of throughput
>  4. Histogram: histogram, support multiple percentage values
>  5. Timer: which can be understood as histogram plus rate
>  ## The measurement data exposure interface only defines the start and stop interfaces
>  # At present, the implementation based on dropwizard and the implementation based on micrometer are provided. The default is the implementation of micrometer, which supports data exposure to Jmx and Prometheus



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