Blindsight

Blindsight is a logging library written in Scala that wraps SLF4J. The name is taken from Peter Watts' excellent first contact novel, Blindsight.

The core feature of Blindsight is that it is "type safe" -- rather than passing in arguments of type java.lang.Object, the API accepts only objects that can be converted into an Argument through the ToArgument type class.

val str: String = "string arg"
val number: Int = 1
val arg: Person = Person(name, age) // has a ToArgument[Person] type class instance
logger.info("{} {} {}", bool, number, person) // compiles fine

logger.info("{}", new Object()) // WILL NOT COMPILE

By adding type safety, Blindsight gives the application more control over how data is logged, rather than implicitly relying on the toString method to render data for logging purposes.

Blindsight adds useful features that solve several outstanding problems with logging:

Rendering structured logs in multiple formats through an AST, along with an optional format-independent DSL.
Providing thread-safe context to logs through context aware logging.
Time-based and targeted logging through conditional logging.
Dynamic targeted logging through scripting.
Easier "printf debugging" through macro based inspections.

Using Scala to break apart the SLF4J API also makes constructing new logging APIs much easier. You have the option of creating your own, depending on your use case:

Building up complex logging statements through fluent logging.
Enforcing user supplied type constraints through semantic logging.
Minimal-overhead tracing and causality tracking through flow logging.
Managing complex relationships and schema through JSON-LD.

Finally, there's also more advanced functionality to transform arguments and statements before entering SLF4J:

Resolving operation-specific loggers through logger resolvers.
Hooks into logging entries through entry transformation
Application accessible debug and trace logs through event buffers

See the documentation for more details.

Blindsight and Echopraxia

If you are looking for a strict structured logging solution in Scala, please checkout echopraxia-plusscala.

Structured logging is optional in Blindsight, and it's possible to mix structured and "flat" arguments and markers into a logging statement. In contrast, echopraxia-plusscala requires structured logging and does not allow unstructured data as input.

Example

You can check out a "starter project" at https://github.com/tersesystems/blindsight-starter.

There's an example application at https://github.com/tersesystems/play-blindsight that integrates with Honeycomb Tracing using the flow logger:

Dependencies

The only hard dependency is the SLF4J API. Structured logging is implemented for Logback with logstash-logback-encoder, but this is only a requirement if you are using structured logging.

Blindsight is a pure SLF4J wrapper: it delegates all logging through to the SLF4J API and does not configure or manage the SLF4J implementation at all.

Versions are published for Scala 2.11, 2.12, 2.13, and 3.0.0.

Install

See Setup for how to install Blindsight.

You can check out a "starter project" at https://github.com/tersesystems/blindsight-starter.

Because Blindsight uses a very recent version of Logstash that depends on Jackson 2.11.0, you may need to update your dependencies for the jackson-scala-module if you're using Play or Akka.

libraryDependencies += "com.fasterxml.jackson.module" %% "jackson-module-scala" % "2.11.0"

Usage

The easiest way to use Blindsight is to import the base package and the DSL:

import com.tersesystems.blindsight._
import com.tersesystems.blindsight.DSL._

To use a Blindsight Logger:

val logger = LoggerFactory.getLogger
logger.info("I am an SLF4J-like logger")

or in block form for diagnostic logging:

logger.debug { debug => debug("I am an SLF4J-like logger") }

Structured DSL:

import com.tersesystems.blindsight._
import com.tersesystems.blindsight.DSL._

logger.info("Logs with argument {}", bobj("array" -> Seq("one", "two", "three")))

Statement Interpolation:

val dayOfWeek = "Monday"
val temp = 72 

// macro expands this to:
// Statement("It is {} and the temperature is {} degrees.", Arguments(dayOfWeek, temp))
val statement: Statement = st"It is ${dayOfWeek} and the temperature is ${temp} degrees."

logger.info(statement)

Marker/Argument Type Classes:

case class Lotto(
  id: Long,
  winningNumbers: List[Int],
  winners: List[Winner],
  drawDate: Option[java.util.Date]
) {
  lazy val asBObject: BObject = "lotto" ->
      ("lotto-id"          -> id) ~
        ("winning-numbers" -> winningNumbers) ~
        ("draw-date"       -> drawDate.map(_.toString)) ~
        ("winners"         -> winners.map(w => w.asBObject))
}

object Lotto {
  implicit val toArgument: ToArgument[Lotto] = ToArgument { lotto => Argument(lotto.asBObject) }
}

val winners =
  List(Winner(23, List(2, 45, 34, 23, 3, 5)), Winner(54, List(52, 3, 12, 11, 18, 22)))
val lotto = Lotto(5, List(2, 45, 34, 23, 7, 5, 3), winners, None)

logger.info("message {}", lotto) // auto-converted to structured output

JSON-LD:

implicit val nodeObjectToArgument: ToArgument[NodeObject] = ToArgument[NodeObject] { nodeObject =>
  Argument(BlindsightASTMapping.toBObject(nodeObject))
}

implicit val nodeObjectToMarkers: ToMarkers[NodeObject] = ToMarkers { nodeObject =>
  Markers(BlindsightASTMapping.toBObject(nodeObject))
}

implicit val nodeObjectToStatement: ToStatement[NodeObject] = ...

class Foo extends LDContext { // LDContext contains all the type safe bindings
  def sayHello(): Unit = {
    val willPerson = NodeObject(
      `@type`    -> "Person",
      `@id`      -> willId,
      givenName  -> "Will",
      familyName -> "Sargent",
      parent     -> parentId,
      occupation -> Occupation(
        estimatedSalary = MonetaryAmount(Currency.getInstance("USD"), 1),
        name = "Code Monkey"
      )
    )

    logger.info("as an argument {}", willPerson) // as an argument
    logger.info(Markers(willPerson), "as a marker") // as a marker
    
    logger.semantic[NodeObject].info(willPerson) // or as a statement
  }
}

Fluent logging:

logger.fluent.info
  .message("The Magic Words are")
  .argument(Arguments("Squeamish", "Ossifrage"))
  .logWithPlaceholders()

Semantic logging:

// log only user events
logger.semantic[UserEvent].info(userEvent)

// Works well with refinement types
import eu.timepit.refined.api.Refined
import eu.timepit.refined.string._
import eu.timepit.refined._
logger.semantic[String Refined Url].info(refineMV(Url)("https://tersesystems.com"))

Flow logging:

import com.tersesystems.blindsight.flow._

implicit def flowBehavior[B]: FlowBehavior[B] = new SimpleFlowBehavior

val arg1: Int = 1
val arg2: Int = 2
val result:Int = logger.flow.trace(arg1 + arg2)

Conditional logging:

logger.withCondition(booleanCondition).info("Only logs when condition is true")

logge

Blindsight

Install / Use

README

Blindsight

Blindsight and Echopraxia

Example

Dependencies

Install

Usage