yeah I think we (Uber) use log4j instead of log4j2

maybe there is some backward compatible version? usually, I see the disk write being the bottleneck and removing that helps a lot as well (use async logger)

if i check-in this, it will pick up the default setting until you write a new log4j2 config and pass it as a jvm option

What is the default setting?

Log4j will provide a default configuration if it cannot locate a configuration file. The default configuration, provided in the DefaultConfiguration class, will set up:

A ConsoleAppender attached to the root logger.
A PatternLayout set to the pattern "%d{HH:mm:ss.SSS} [%t] %-5level %logger{36} - %msg%n" attached to the ConsoleAppender
Note that by default Log4j assigns the root logger to Level.ERROR.

The output of MyApp would be similar to:

17:13:01.540 [main] ERROR com.foo.Bar - Did it again!
17:13:01.540 [main] ERROR MyApp - Didn't do it.

https://logging.apache.org/log4j/2.x/manual/configuration.html

it will output the log to console, set the level to

error

, and the format is like the above example

just curious have we considered using the log4j 1->2 bridge/compatible library instead of direct migration, such as this: https://logging.apache.org/log4j/log4j-2.2/log4j-1.2-api/index.html

@snlee ^^

@User ^^

Thanks @User for correcting

If you check the pr, we do import log4j 1->2 bridge. But, this is for routing log4j 1.x apis to log4j2. This is more for making the code backward compatible, not the config (if you coded your application using log4j 1.x api, importing this will redirect log4j1.x api to log4j2). In our case, we use slf4j for logging api so it doesn’t apply but I pulled that bridge because some of our dependencies are directly using log4j 1.x api.

@User ^^

If this code goes to the official apache release to 0.2.0, we will include this information to the release note.

Hello everyone! I was wondering if anyone ran into controller.out taking up disk space issue. Pinot controller logs don't seem to clear or rotate by default. Any suggestions on how to set that? thank you!

Hello Everyone ! Has anyone added their own jdbc datasource to third eye

?

Will respond in few hours

@User @User, Thanks for the revert. few insights: 1. Currently app/#/self-serve/import-sql-metric is not reachable directly from create alert page. 2. mysqlDBNameToDataSourceMap in SqlResponseCacheLoader is not populated with datasource and is null.

I have added redshift jdbc connection like this

there is another slack channel for thirdeye

👍

<!here> I merged log4j2 migration commit. If you are deploying after this change, you will need to rewrite log4j configuration file to log4j2 format. We will update note on this for the next apache release (0.2.0). https://github.com/apache/incubator-pinot/pull/4139

Hey folks, is there any thinking around automatically discarding "null" (default) values when querying Pinot. I think the current guidance we give to customers is that filter out explicitly in your query which is a bit inconvenient.

you mean support for NULL ?

yes

Hey guys, I am trying to import a metric to ThirdEye that is calculated by dividing two numbers (a percentage). I can't create a pinot column that has that value calculated as percentage is not additive and breakdown by dimensions will not be possible. Is there a way to define a metric in ThirdEye that is calculated by combining other metrics on the fly?

A concrete example: day 1 retention (number of users who logged in today and registered yesterday / total number of users registered yesterday). I would like to create alerts on that metric and also be able to break down by all dimensions in root cause analysis

Yes. There is concept of derived metric and you can setup the expression to compute it on the admin UI/rest api.

Perfect, thanks! Are derived metrics documented anywhere? Most of it is clear, but for example no idea what is Inverse metric supposed to be?