See https://petabridge.com/blog/you-dont-need-akka-healthchecks-anymore/ - all of the functionality from Akka.HealthChecks is now built-in to Akka.Hosting, with fewer bugs and configuration overhead.
The core plugin is designed to run as standalone health check endpoint for your actor system. Three types of transports are supported:
Custom (default)
Used to specify that no built-in transport will be used. Typically users will query / subscribe to the Readiness or Liveness probe actors and pipe the changes in liveness / readiness status out to something like a custom HTTP endpoint.
TCP socket
This transport mode fulfills a Kubernetes-like style health check endpoint, where liveness/readiness are tested by checking if a pod can accept a socket connection request on a certain port. See the Kubernetes documentation for usage example.
File
Writes the readiness or liveness status out to disk to a specified file location. Used in combination with liveness checks such as "command line execution" checks. See the Kubernetes documentation for usage example.
There are 5 probe providers that can be used with Akka.HealthCheck:
DefaultReadinessProvider - The default readiness probe. The probe reports the time the ActorSystem was started.
Available inside the Akka.HealthCheck NuGet package.
DefaultLivenessProvider - The default liveness probe. The probe reports the time the ActorSystem was started.
Available inside the Akka.HealthCheck NuGet package.
ClusterReadinessProbeProvider - Readiness probe for clustering.
Reports unhealthy when:
All other nodes in the cluster is unreachable.
Available inside the Akka.HealthCheck.Cluster NuGet package.
ClusterLivenessProbeProvider - Liveness probe for clustering.
Reports unhealthy when:
The ActorSystem left the cluster.
Available inside the Akka.HealthCheck.Cluster NuGet package.
AkkaPersistenceLivenessProbeProvider - Liveness probe for persistence. It probes the persistence storage every second to check that persistence is working properly.
Reports healthy when persistence successfully recover both journal and snapshot data from storage.
Reports unhealthy when:
Persistence just started and has not recovered.
Either journal or snapshot failed recovery inside the periodic check.
Available inside the Akka.HealthCheck.Persistence NuGet package.
To use the Akka.Hosting extension method, you will need to install the Akka.HealthCheck.Hosting package.
dotnet add package Akka.HealthCheck.Hosting
NOTE
Unlike the core library, Akka.HealthCheck.Hosting already includes Akka.HealthCheck.Cluster and Akka.HealthCheck.Persistence as its dependency; there is no need to install these packages to start using Akka.HealthCheck with clustering or persistence.
To add health check to your actor system, use the .WithHealthCheck() hosting extension method:
using var host = new HostBuilder()
.ConfigureServices((hostContext, services) =>
{
services.AddAkka("test-system", (builder, provider) =>
{
// Add Akka.Cluster support
builder.WithClustering();
// Add persistence
builder
.WithInMemoryJournal()
.WithInMemorySnapshotStore();
// Add Akka.HealthCheck
builder.WithHealthCheck(options =>
{
// Here we're adding all of the built-in providers
options.AddProviders(HealthCheckType.All);
});
});
})
.Build();
await host.RunAsync();
HealthCheckType is a bit flag enum that consists of these choices:
The cluster and persistence probe providers for Akka.HealthCheck are published in separate NuGet packages.
If you need to use the cluster liveness and readiness probe, you need to install the Akka.HealthCheck.Cluster NuGet package
If you need to use the persistence liveness probe, you need to install the Akka.HealthCheck.Persistence NuGet package
Akka.HealthCheck can be added manually through HOCON configuration:
var hocon = @"
akka.healthcheck {
log-config-on-start = on
log-info = on
liveness {
providers {
default = "Akka.HealthCheck.Liveness.DefaultLivenessProvider, Akka.HealthCheck"
cluster = "Akka.HealthCheck.Cluster.ClusterLivenessProbeProvider, Akka.HealthCheck.Cluster"
}
transport = tcp
tcp.port = 8080
}
readiness {
providers {
default = "Akka.HealthCheck.Readiness.DefaultReadinessProvider, Akka.HealthCheck"
custom = "MyAssembly.CustomReadinessProvider, MyAssembly"
}
transport = file
file.path = ""snapshot.txt""
}
}";
var config = ConfigurationFactory.ParseString(hocon);
var actorSystem = ActorSystem.Create("Probe", config);
var healthCheck = AkkaHealthCheck.For(actorSystem);
actorSystem.WhenTerminated.Wait();
The full reference HOCON configuration are:
############################################
# Akka.HealthCheck Reference Config File #
############################################
akka.healthcheck{
# Log the complete configuration at INFO level when the actor system is started.
# This is useful when you are uncertain of what configuration is used.
log-config-on-start = on
# Log Liveness and Readiness probe event messages
# Such as Liveness/Readiness subscriptions, and status request
log-info = on
liveness {
# List of liveness probe providers.
# Custom end-user provider can be created by implementing the IProbeProvider interface.
providers {
# The default IProbeProvider implementation used for executing
# liveness checks inside Akka.HealthCheck
default = "Akka.HealthCheck.Liveness.DefaultLivenessProvider, Akka.HealthCheck"
# Clustering liveness check provider.
# To use, install the Akka.HealthCheck.Cluster NuGet package and uncomment this line.
#
#cluster = "Akka.HealthCheck.Cluster.ClusterLivenessProbeProvider, Akka.HealthCheck.Cluster"
# Persistence liveness check provider.
# To use, install the Akka.HealthCheck.Persistence NuGet package and uncomment this line.
#
#persistence = "Akka.HealthCheck.Persistence.AkkaPersistenceLivenessProbeProvider, Akka.HealthCheck.Persistence"
}
# Defines the interval for each persistence health check probe refresh
persistence.probe-interval = 10s
# Defines the signaling mechanism used to communicate with K8s, AWS, Azure,
# or whatever the hosting environment is for the Akka.NET application. The
# accepted values are 'file', 'tcp', and 'custom'.
#
# In the event of a custom transport (which is the default), Akka.HealthCheck
# won't try to automatically report any probe data to any medium. It's up
# to the end-user to query that data directly from the AkkaHealthCheck
# ActorSystem extension.
transport = custom
# If the `transport` used is `file`, this is where we specify the path of the file
# that we will write status data to. It's strongly recommended that you use an
# absolute path for best results.
file.path = "liveness.txt"
# If the `transport` used is `tcp`, this is where we specify the port # of the inbound
# socket that we're going open in order to accept external healthcheck connections.
tcp.port = 11000
}
readiness{
providers {
# The default IProbeProvider implementation used for executing
# readiness checks inside Akka.Healtcheck. Can be overridden by
# end-users via a custom IProbeProvider implementation.
default = "Akka.HealthCheck.Readiness.DefaultReadinessProvider, Akka.HealthCheck"
# Clustering readiness check provider.
# To use, install the Akka.HealthCheck.Cluster NuGet package and uncomment this line.
#
#cluster = "Akka.HealthCheck.Cluster.ClusterReadinessProbeProvider, Akka.HealthCheck.Cluster"
}
# Defines the signaling mechanism used to communicate with K8s, AWS, Azure,
# or whatever the hosting environment is for the Akka.NET application. The
# accepted values are 'file', 'tcp', and 'custom'.
#
# In the event of a custom transport (which is the default), Akka.HealthCheck
# won't try to automatically report any probe data to any medium. It's up
# to the end-user to query that data directly from the AkkaHealthCheck
# ActorSystem extension.
transport = custom
# If the `transport` used is `file`, this is where we specify the path of the file
# that we will write status data to. It's strongly recommended that you use an
# absolute path for best results.
file.path = "readiness.txt"
# If the `transport` used is `tcp`, this is where we specify the port # of the inbound
# socket that we're going open in order to accept external healthcheck connections.
tcp.port = 11001
}
}
Akka.HealthCheck and Akka.Hosting can be directly integrated with Microsoft.AspNetCore.Diagnostics.HealthChecks, allowing users to access Akka.HealthCheck via HTTP REST API.
The package provides 5 IHealthCheck probes that can be registered with the health check middleware, each uniquely tagged so they can be individually filtered during mapping:
AkkaReadinessProbe - The default readiness probe. The probe reports the time the ActorSystem was started.
Tags: [ "akka", "ready", "node" ]
AkkaLivenessProbe - The default liveness probe. The probe reports the time the ActorSystem was started.
Tags: [ "akka", "live", "node" ]
AkkaClusterReadinessProbe - Readiness probe for clustering.
Reports healthy when:
The ActorSystem joined a cluster.
The ActorSystem is connected to a cluster
Reports unhealthy when:
The ActorSystem just started has not joined a cluster.
All other nodes in the cluster is unreachable.
Tags: [ "akka", "ready", "cluster" ]
AkkaClusterLivenessProbe - Liveness probe for clustering.
Reports healthy when:
The ActorSystem joined a cluster.
The ActorSystem is connected to a cluster
Reports unhealthy when:
The ActorSystem just started and has not joined a cluster.
The ActorSystem left the cluster.
Tags: [ "akka", "live", "cluster" ]
AkkaPersistenceLivenessProbe - Liveness probe for persistence. It probes the persistence storage every second to check that persistence is working properly.
Reports healthy when persistence successfully recover both journal and snapshot data from storage.
Reports unhealthy when:
Persistence just started and has not recovered.
Either journal or snapshot failed recovery inside the periodic check.
The convenience IServiceCollection extension method WithAkkaHealthCheck(HealthCheckType) can be used to register the standard probes in any combination.
var webBuilder = WebApplication.CreateBuilder(args);
webBuilder.Services
.WithAkkaHealthCheck(HealthCheckType.All);
As alternative, individual probes can be registered using these methods:
The convenience AkkaConfigurationBuilder extension method WithWebHealthCheck(IServiceProvider) automatically scans for any registered probes inside the health check middleware and adds the respective Akka.NET health check probes to the ActorSystem
var webBuilder = WebApplication.CreateBuilder(args);
webBuilder.Services
.WithAkkaHealthCheck(HealthCheckType.All)
.AddAkka("actor-system", (builder, serviceProvider) =>
{
// Automatically detects which health checks were registered
// inside the health check middleware and starts them
builder.WithWebHealthCheck(serviceProvider);
});
The convenience IEndpointRouteBuilder extension method MapAkkaHealthCheckRoutes automatically scans for any registered probes inside the health check middleware and maps all the probes to a HTTP route. The HTTP route is the concatenation of the probe tags. By default:
AkkaReadinessProbe is mapped to "/healthz/akka/ready/node"
AkkaLivenessProbe is mapped to "/healthz/akka/live/node"
AkkaClusterReadinessProbe is mapped to "/healthz/akka/ready/cluster"
AkkaClusterLivenessProbe is mapped to "/healthz/akka/live/cluster"
AkkaPersistenceLivenessProbe is mapped to "/healthz/akka/live/persistence"
All liveness probes can be queried all at once at "/healthz/akka/live"
All readiness probes can be queried all at once at "/healthz/akka/ready"
All Akka.NET probes can be queried all at once at "/healthz/akka"
var webBuilder = WebApplication.CreateBuilder(args);
webBuilder.Services
// Register all of the health check service with IServiceCollection
.WithAkkaHealthCheck(HealthCheckType.All)
.AddAkka("actor-system", (builder, serviceProvider) =>
{
builder
// Automatically detects which health checks were registered
// inside the health check middleware and starts them
.WithWebHealthCheck(serviceProvider);
});
var app = webBuilder.Build();
// Automatically detects which health checks were registered inside
// the health check middleware and maps their routes
app.MapAkkaHealthCheckRoutes();
await app.RunAsync();
By default, the health check middleware outputs a simple string response of either "healthy" or "unhealthy" regardless of the number of probes being queried. To more verbose response can be gained by using Helper.JsonResponseWriter as the route endpoint response writer.
app.MapAkkaHealthCheckRoutes(
optionConfigure: (_, opt) =>
{
// Use a custom response writer to output a json of all reported statuses
opt.ResponseWriter = Helper.JsonResponseWriter;
});
The attached build script will automatically do the following based on the conventions of the project names added to this project:
Any project name ending with .Tests will automatically be treated as a XUnit2 project and will be included during the test stages of this build script;
Any project name ending with .Tests will automatically be treated as a NBench project and will be included during the test stages of this build script; and
Any project meeting neither of these conventions will be treated as a NuGet packaging target and its .nupkg file will automatically be placed in the bin\nuget folder upon running the build.[cmd|sh] all command.
This solution also supports DocFx for generating both API documentation and articles to describe the behavior, output, and usages of your project.
All of the relevant articles you wish to write should be added to the /docs/articles/ folder and any API documentation you might need will also appear there.
All of the documentation will be statically generated and the output will be placed in the /docs/_site/ folder.
To preview the documentation for this project, execute the following command at the root of this folder:
C:\> serve-docs.cmd
This will use the built-in docfx.console binary that is installed as part of the NuGet restore process from executing any of the usual build.cmd or build.sh steps to preview the fully-rendered documentation. For best results, do this immediately after calling build.cmd buildRelease.
This project will automatically populate its release notes in all of its modules via the entries written inside RELEASE_NOTES.md and will automatically update the versions of all assemblies and NuGet packages via the metadata included inside Directory.Build.props.
All new projects added into the solution will automatically picks up all the settings set inside Directory.Build.props.
This project uses SignService to code-sign NuGet packages prior to publication. The build.cmd and build.sh scripts will automatically download the SignClient needed to execute code signing locally on the build agent, but it's still your responsibility to set up the SignService server per the instructions at the linked repository.
Once you've gone through the ropes of setting up a code-signing server, you'll need to set a few configuration options in your project in order to use the SignClient: