Files
ECMJobRunner/ECMJobRunner.WebCron/ProfileWorker/ProfileWorker.cs
TekH 8a05d86285 feat: Add health check monitoring for ProfileWorker
ProfileWorker changes:
- Implement IHealthCheck interface
- Replace _workCount with DateTime-based tracking
- Track _lastSuccessfulRun, _consecutiveFailures, _lastException
- Graceful shutdown handling (OperationCanceledException)
- Health states: Healthy, Degraded (1-2 failures), Unhealthy (3x interval)

DependencyInjection changes:
- Register ProfileWorker as singleton (for health check access)
- Use factory pattern for IHostedService registration

Program.cs changes:
- Add health check service with ProfileWorker
- Map /health endpoint (full JSON response with all checks)
- Map /health/ready endpoint (filtered by 'ready' tag)
- Custom JSON response writer with detailed metrics
2026-07-13 12:19:11 +02:00

120 lines
4.2 KiB
C#

using Microsoft.Extensions.Diagnostics.HealthChecks;
using Microsoft.Extensions.Options;
namespace ECMJobRunner.WebCron.ProfileWorker;
public class ProfileWorker(
ILogger<ProfileWorker> Logger,
IServiceScopeFactory ScopeFactory,
IOptions<ProfileWorkerOptions> Options) : BackgroundService, IHealthCheck
{
private readonly ProfileWorkerOptions _options = Options.Value;
// Health check state
private DateTime _lastSuccessfulRun = DateTime.UtcNow;
private int _consecutiveFailures = 0;
private Exception? _lastException = null;
protected override async Task ExecuteAsync(CancellationToken stoppingToken)
{
Logger.LogInformation("ProfileWorker started");
while (!stoppingToken.IsCancellationRequested)
{
try
{
// Create a scope to resolve scoped services (ISQLExecutor used by MediatR pipeline)
using var scope = ScopeFactory.CreateScope();
var work = scope.ServiceProvider.GetRequiredService<ProfileWork>();
await work.ExecuteAsync(stoppingToken);
// Success - update health state
_lastSuccessfulRun = DateTime.UtcNow;
_consecutiveFailures = 0;
_lastException = null;
if (Logger.IsEnabled(LogLevel.Debug))
{
Logger.LogDebug("ProfileWorker sync completed successfully at {time}", _lastSuccessfulRun);
}
}
catch (OperationCanceledException) when (stoppingToken.IsCancellationRequested)
{
// Graceful shutdown - app is stopping
Logger.LogInformation("ProfileWorker stopping due to cancellation request");
break;
}
catch (Exception ex)
{
// Unexpected error - track for health check
_consecutiveFailures++;
_lastException = ex;
Logger.LogError(ex,
"An unexpected error occurred in ProfileWorker (consecutive failures: {failures})",
_consecutiveFailures);
}
await Task.Delay(_options.IntervalMS, stoppingToken);
}
Logger.LogInformation("ProfileWorker stopped");
}
public Task<HealthCheckResult> CheckHealthAsync(
HealthCheckContext context,
CancellationToken cancellationToken = default)
{
var timeSinceLastSuccess = DateTime.UtcNow - _lastSuccessfulRun;
var maxAllowedDelay = TimeSpan.FromMilliseconds(_options.IntervalMS * 3);
// Unhealthy: No successful run for 3x interval
if (timeSinceLastSuccess > maxAllowedDelay)
{
var data = new Dictionary<string, object>
{
["LastSuccessfulRun"] = _lastSuccessfulRun,
["TimeSinceLastSuccess"] = timeSinceLastSuccess,
["ConsecutiveFailures"] = _consecutiveFailures,
["IntervalMS"] = _options.IntervalMS
};
if (_lastException != null)
{
data["LastException"] = _lastException.Message;
}
return Task.FromResult(HealthCheckResult.Unhealthy(
$"ProfileWorker has not completed successfully for {timeSinceLastSuccess.TotalSeconds:F0} seconds ({_consecutiveFailures} consecutive failures)",
_lastException,
data
));
}
// Degraded: 1-2 consecutive failures but within time limit
if (_consecutiveFailures > 0)
{
var data = new Dictionary<string, object>
{
["LastSuccessfulRun"] = _lastSuccessfulRun,
["ConsecutiveFailures"] = _consecutiveFailures
};
return Task.FromResult(HealthCheckResult.Degraded(
$"ProfileWorker has {_consecutiveFailures} consecutive failure(s) but still operational",
null,
data
));
}
// Healthy
return Task.FromResult(HealthCheckResult.Healthy(
"ProfileWorker is running normally",
new Dictionary<string, object>
{
["LastSuccessfulRun"] = _lastSuccessfulRun
}
));
}
}