Healthie.NET.Checkers 4.1.3

Healthie.NET.Checkers

▶ Live demo — board.healthie-dotnet.dev — a read-only Healthie.NET dashboard watching real status pages (Anthropic, OpenAI, GitHub, Cloudflare, and more), built from these packages. Full documentation at healthie-dotnet.dev.

Ready-made pulse checkers for the things nearly every application ends up watching, so there is no checker code to write.

Installation

dotnet add package Healthie.NET.Checkers

Usage

using Healthie.Abstractions.Scheduling;
using Healthie.Checkers;

builder.Services
    .AddHealthie(typeof(Program).Assembly)
    .AddHealthieHttpChecker("payments-api", new Uri("https://payments.internal/health"),
        PulseSchedule.Every(TimeSpan.FromSeconds(30)))
    .AddHealthieTcpChecker("postgres", "db.internal", 5432,
        PulseSchedule.Every(TimeSpan.FromMinutes(1)))
    .AddHealthieCertificateChecker("public-tls", "example.com",
        PulseSchedule.Cron("0 3 * * *"), warnWithin: TimeSpan.FromDays(30))
    .AddHealthieDnsChecker("apex-dns", "example.com",
        PulseSchedule.Every(TimeSpan.FromMinutes(5)))
    .AddHealthieDiskSpaceChecker("data-volume", "/data",
        PulseSchedule.Cron("*/15 * * * *"));

Each call takes a name. The same type is usually registered several times — three endpoints, two drives — and the name is what separates them in storage, in the REST API and on the dashboard.

Unlike checkers you write yourself, these are not found by assembly scanning: each has to be told what to watch.

The checkers

Two of them use all three states

Certificate expiry and disk space are gradual and knowable in advance, so both report suspicious before they report unhealthy — a certificate inside its warning window, a drive below the warning threshold but above the critical one. That warning is the useful signal; by the time either is unhealthy it is already an outage.

The others are binary and only ever report healthy or unhealthy. Give them an unhealthyThreshold if you want a blip to read as suspicious rather than an outage.

HTTP client configuration

The HTTP checker resolves its client from IHttpClientFactory under the name HttpPulseChecker.HttpClientName. Configure its timeout, handler or retry policy by naming it in your own registration:

builder.Services.AddHttpClient(HttpPulseChecker.HttpClientName, client =>
{
    client.Timeout = TimeSpan.FromSeconds(5);
});

The request reads headers only — a health endpoint that returns a body is common, and downloading it to discard it is work this check does not need to do.