Skip to main content
Version: ZIO 1.x

Runtime

A Runtime[R] is capable of executing tasks within an environment R.

To run an effect, we need a Runtime, which is capable of executing effects. Runtimes bundle a thread pool together with the environment that effects need.

What is a Runtime System?

Whenever we write a ZIO program, we create a ZIO effect from ZIO constructors plus using its combinators. We are building a blueprint. ZIO effect is just a data structure that describes the execution of a concurrent program. So we end up with a tree data structure that contains lots of different data structures combined together to describe what the ZIO effect should do. This data structure doesn't do anything, it is just a description of a concurrent program.

So the most thing we should keep in mind when we are working with a functional effect system like ZIO is that when we are writing code, printing a string onto the console, reading a file, querying a database, and so forth; We are just writing a workflow or blueprint of an application. We are just building a data structure.

So how ZIO run these workflows? This is where ZIO Runtime System comes into play. Whenever we run an unsaferun function, the Runtime System is responsible to step through all the instructions described by the ZIO effect and execute them.

To simplify everything, we can think of a Runtime System like a black box that takes both the ZIO effect (ZIO[R, E, A]) and its environment (R), it will run this effect and then will return its result as an Either[E, A] value.

ZIO Runtime System

Responsibilities of the Runtime System

Runtime Systems have a lot of responsibilities:

  1. Execute every step of the blueprint โ€” They have to execute every step of the blueprint in a while loop until it's done.

  2. Handle unexpected errors โ€” They have to handle unexpected errors, not just the expected ones but also the unexpected ones.

  3. Spawn concurrent fiber โ€” They are actually responsible for the concurrency that effect systems have. They have to spawn a fiber every time we call fork on an effect to spawn off a new fiber.

  4. Cooperatively yield to other fibers โ€” They have to cooperatively yield to other fibers so that fibers that are sort of hogging the spotlight, don't get to monopolize all the CPU resources. They have to make sure that the fibers split the CPU cores among all the fibers that are working.

  5. Capture execution and stack traces โ€” They have to keep track of where we are in the progress of our own user-land code so the nice detailed execution traces can be captured.

  6. Ensure finalizers are run appropriately โ€” They have to ensure finalizers are run appropriately at the right point in all circumstances to make sure that resources are closed that clean-up logic is executed. This is the feature that powers ZManaged and all the other resource-safe constructs in ZIO.

  7. Handle asynchronous callback โ€” They have to handle this messy job of dealing with asynchronous callbacks. So we don't have to deal with async code. When we are doing ZIO, everything is just async out of the box.

Running a ZIO Effect

There are two ways to run ZIO effect:

  1. Using zio.App entry point
  2. Using unsafeRun method directly

Using zio.App

In most cases we use this method to run our ZIO effect. zio.App has a run function which is the main entry point for running a ZIO application on the JVM:

package zio
trait App {
def run(args: List[String]): URIO[ZEnv, ExitCode]
}

Assume we have written an effect using ZIO:

import zio.console._

def myAppLogic =
for {
_ <- putStrLn("Hello! What is your name?")
n <- getStrLn
_ <- putStrLn("Hello, " + n + ", good to meet you!")
} yield ()

Now we can run that effect using run entry point:

object MyApp extends zio.App {
final def run(args: List[String]) =
myAppLogic.exitCode
}

Using unsafeRun

Another way to execute ZIO effect is to feed the ZIO effect to the unsafeRun method of Runtime system:

object RunZIOEffectUsingUnsafeRun extends scala.App {
zio.Runtime.default.unsafeRun(
myAppLogic
)
}

We don't usually use this method to run our effects. One of the use cases of this method is when we are integrating the legacy (non-effectful code) with the ZIO effect. It also helps us to refactor a large legacy code base into a ZIO effect gradually; Assume we have decided to refactor a component in the middle of a legacy code and rewrite that with ZIO. We can start rewriting that component with the ZIO effect and then integrate that component with the existing code base, using the unsafeRun function.

Default Runtime

ZIO contains a default runtime called Runtime.default, configured with the ZEnv (the default ZIO environment) and a default Platform designed to work well for mainstream usage. It is already implemented as below:

object Runtime {
lazy val default: Runtime[ZEnv] = Runtime(ZEnv.Services.live, Platform.default)
}

The default runtime includes a default Platform which contains minimum capabilities to bootstrap execution of ZIO tasks and live (production) versions of all ZIO built-in services. The default ZIO environment (ZEnv) for the JS platform includes Clock, Console, System, Random; and the JVM platform also has a Blocking service:

// Default JS environment
type ZEnv = Clock with Console with System with Random

// Default JVM environment
type ZEnv = Clock with Console with System with Random with Blocking

We can easily access the default Runtime to run an effect:

object MainApp extends scala.App {
val runtime = Runtime.default
runtime.unsafeRun(myAppLogic)
}

Custom Runtime

Sometimes we need to create a custom Runtime with a user-defined environment and user-specified Platform. Many real applications should not use Runtime.default. Instead, they should make their own Runtime which configures the Platform and environment accordingly.

Some use-cases of custom Runtimes:

Providing Environment to Runtime System

The custom runtime can be used to run many different effects that all require the same environment, so we don't have to call ZIO#provide on all of them before we run them.

For example, assume we want to create a Runtime for services that are for testing purposes, and they don't interact with real external APIs. So we can create a runtime, especially for testing.

Let's say we have defined two Logging and Email services:

trait Logging {
def log(line: String): UIO[Unit]
}

object Logging {
def log(line: String): URIO[Has[Logging], Unit] =
ZIO.serviceWith[Logging](_.log(line))
}

trait Email {
def send(user: String, content: String): Task[Unit]
}

object Email {
def send(user: String, content: String): ZIO[Has[Email], Throwable, Unit] =
ZIO.serviceWith[Email](_.send(user, content))
}

We are going to implement a live version of Logging service and also a mock version of Email service for testing:

case class LoggingLive() extends Logging {
override def log(line: String): UIO[Unit] =
ZIO.effectTotal(print(line))
}

case class EmailMock() extends Email {
override def send(user: String, content: String): Task[Unit] =
ZIO.effect(println(s"sending email to $user"))
}

Let's create a custom runtime that contains these two service implementations in its environment:

val testableRuntime = Runtime(
Has.allOf[Logging, Email](LoggingLive(), EmailMock()),
Platform.default
)

Also, we can map the default runtime to the new runtime, so we can append new services to the default ZIO environment:

val testableRuntime: Runtime[zio.ZEnv with Has[Logging] with Has[Email]] =
Runtime.default
.map((zenv: zio.ZEnv) =>
zenv ++ Has.allOf[Logging, Email](LoggingLive(), EmailMock())
)

Now we can run our effects using this custom Runtime:

testableRuntime.unsafeRun(
for {
_ <- Logging.log("sending newsletter")
_ <- Email.send("David", "Hi! Here is today's newsletter.")
} yield ()
)

Application Monitoring

Sometimes to diagnose runtime issues and understand what is going on in our application we need to add some sort of monitoring task to the Runtime System. It helps us to track fibers and their status.

By adding a Supervisor to the current platform of the Runtime System, we can track the activity of fibers in a program. So every time a fiber gets started, forked, or every time a fiber ends its life, all these contextual pieces of information get reported to that Supervisor.

For example, the ZIO ZMX enables us to monitor our ZIO application. To include that in our project we must add the following line to our build.sbt:

libraryDependencies += "dev.zio" %% "zio-zmx" % "0.0.6"

ZIO ZMX has a specialized Supervisor called ZMXSupervisor that can be added to our existing Runtime:

import zio._
import zio.console._
import zio.zmx._
import zio.zmx.diagnostics._

val program: ZIO[Console, Throwable, Unit] =
for {
_ <- putStrLn("Waiting for input")
a <- getStrLn
_ <- putStrLn("Thank you for " + a)
} yield ()

val diagnosticsLayer: ZLayer[ZEnv, Throwable, Has[Diagnostics]] =
Diagnostics.make("localhost", 1111)

val runtime: Runtime[ZEnv] =
Runtime.default.mapPlatform(_.withSupervisor(ZMXSupervisor))

runtime.unsafeRun(program.provideCustomLayer(diagnosticsLayer))

Application Tracing

We can enable or disable execution tracing or configure its setting. Execution tracing has full of junk. There are lots of allocations that all need to be garbage collected afterward. So it has a tremendous impact on the complexity of the application runtime.

Users often turn off tracing in critical areas of their application. Also, when we are doing benchmark operation, it is better to create a Runtime without tracing capability:

import zio.internal.Tracing
import zio.internal.tracing.TracingConfig

val rt1 = Runtime.default.mapPlatform(_.withTracing(Tracing.disabled))
val rt2 = Runtime.default.mapPlatform(_.withTracing(Tracing.enabledWith(TracingConfig.stackOnly)))

val config = TracingConfig(
traceExecution = true,
traceEffectOpsInExecution = true,
traceStack = true,
executionTraceLength = 100,
stackTraceLength = 100,
ancestryLength = 10,
ancestorExecutionTraceLength = 10,
ancestorStackTraceLength = 10
)
val rt3 = Runtime.default.mapPlatform(_.withTracingConfig(config))

User-defined Executor

An executor is responsible for executing effects. The way how each effect will be run including detail of threading, scheduling, and so forth, is separated from the caller. So, if we need to have a specialized executor according to our requirements, we can provide that to the ZIO Runtime:

import zio.internal.Executor
import java.util.concurrent.{ThreadPoolExecutor, TimeUnit, LinkedBlockingQueue}

val runtime = Runtime.default.mapPlatform(
_.withExecutor(
Executor.fromThreadPoolExecutor(_ => 1024)(
new ThreadPoolExecutor(
5,
10,
5000,
TimeUnit.MILLISECONDS,
new LinkedBlockingQueue[Runnable]()
)
)
)
)

Benchmarking

To do benchmark operation, we need a Runtime with settings suitable for that. It would be better to disable tracing and auto-yielding. ZIO has a built-in Platform proper for benchmark operations, called Platform.benchmark which we can map the default Platform to the benchmark version:

val benchmarkRuntime = Runtime.default.mapPlatform(_ => Platform.benchmark)