Conditions are objects that evaluate entries for acceptance. They are an important part of a timber configuration graph. The result of a condition’s evaluation is ternary:

You can either take advantage of the built-in conditions, which will probably meet most of your needs, or create your own.

Built-in Conditions

There are currently four types of conditions built-in to timber:

Int Conditions

Int conditions support the following operations:

Hopefully, these are self-explanatory.

The use of === versus is is entirely up to you. They behave exactly the same. Choosing one over the other depends on whether you’d rather be able to read your timber configuration like English or rather spend hours trying to figure out that you only put two equals signs in your configuration instead of three. :)


The only Int field currently built into timber is the level field. You can use it to create a condition based on the level of the entry. If the entry does not have a level (entry.level == None), the condition will abstain. The RHS of a level condition operator must be a Level. It can be one of the standard levels or a level that you’ve defined yourself. You can also specify the comparison level using an Int and it will be converted to a Level for you. All comparisons are done based on the integer value of the levels. The names are completely ignored.


import org.scalawag.timber.api.level.Level._
import org.scalawag.timber.backend.dispatcher.configuration.dsl._

level > 5
level >= INFO
level is ERROR

String Conditions

String conditions support the following operations:

The argument for each is a pattern that can be specified either as a String, a scala.Regex or a java.util.Pattern. The operators probably do exactly what they sound like they do.

Fields that are currently available to create String conditions are:

All of these are optional (may not be present in the entry) except for If the field does not have a value for a given entry, all conditions will abstain when evaluating the entry.


import org.scalawag.timber.backend.dispatcher.configuration.dsl._

loggingClass startsWith "org.scalawag.timber"
message contains "[Ee][Rr][Rr][Oo][Rr]".r
logger("clientIpAddress") is "" endsWith "-test"
thread("subsystem").any is "db

Tag Conditions

Tag conditions are not so much a group as a single condition. You can use a Tag condition to determine if an entry has been tagged with a specific Tag.


import org.scalawag.timber.backend.dispatcher.configuration.dsl._

object MyTag extends Tag


Logical Conditions

In addition to the above conditions, timber also allows you to combine conditions to make more complex decisions using logical operations on other conditions. The logical operations that timber supports are:

Again, they should do what you expect them to. The only thing that may need clarification is that if any operand of a logical condition abstains, the logical condition itself abstains. The English words and operators are equivalent. They should operate exactly the same. Which one to use is up to you, based on your style preferences. Scala’s operator precedence may also help you decide on one versus the other, as the rules are different for symbols versus letters.


import org.scalawag.timber.backend.dispatcher.configuration.dsl._

object AlertTag extends Tag

! tagged(AlertTag)
( level > INFO ) or ( loggingClass startsWith "org.scalawag" )

Custom Conditions

From Scratch

Conditions are represented by the org.scalawag.timber.backend.dispatcher.configuration.dsl.Condition trait. The single abstract method it declares is:

def accepts(entry:EntryFacets):Option[Boolean]

This method should return one of the following three responses.

It’s important that you handle the abstain case (None) properly because it’s used internally by timber to optimize the configuration graph. It may pass a hypothetical EntryFacets to your condition prior to entry evaluation during dispatch to determine whether it can prune or bypass a path in the DAG.

Piggy-back Implementations

If you just want to create Int or String conditions based on a field that doesn’t currently have an extractor in timber, you can piggy-back on some of the existing timber code both to make your job easier and to ensure that it behaves the same as other timber condition fields of the same type (i.e., supporting the same comparison operators).

Here’s an example of using timber’s IntConditionFactory to create a new field that contains the depth of a thread attribute stack.

val loggerAttributeCount = IntConditionFactory("loggerAttributeCount") { entry =>

All you have to provide is an extractor function that extracts your field’s value from a partial entry. This function should return None if the field is absent from the entry and Some if it has a definite value. Note that this condition field is not very useful. That’s why it’s not built into timber!

Help improve this page.