Class TelemetryConfig

Namespace

AiDotNet.Deployment.Configuration

Assembly

AiDotNet.dll

Configuration for telemetry - tracking and monitoring model inference metrics.

public class TelemetryConfig

Inheritance

object

TelemetryConfig

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

For Beginners: Telemetry collects data about how your model is performing in production. Think of it like a fitness tracker for your AI model - it tracks important health metrics so you know when something goes wrong or when performance degrades.

What gets tracked:

• Latency: How long each inference takes (helps identify slowdowns)
• Throughput: How many inferences per second (measures capacity)
• Errors: When predictions fail (helps identify issues)
• Cache hits/misses: How often cached models are used (optimizes memory)
• Version usage: Which model versions are being used (helps with rollouts)

Why it's important:

• Detect performance degradation before users complain
• Understand usage patterns to optimize resources
• Debug production issues with real data
• Make data-driven decisions about model updates

Privacy Note: Telemetry doesn't collect user data or predictions, only metadata like timing and version information.

Properties

CollectErrors

Alias for TrackErrors for more intuitive access.

public bool CollectErrors { get; set; }

Property Value

bool

CollectLatency

Alias for TrackLatency for more intuitive access.

public bool CollectLatency { get; set; }

Property Value

bool

CustomTags

Gets or sets custom tags to include with all telemetry events.

public Dictionary<string, string> CustomTags { get; set; }

Property Value

Dictionary<string, string>

Remarks

For Beginners: Add custom labels to telemetry data (e.g., environment, region). Useful for filtering and organizing metrics from different deployments.

Enabled

Gets or sets whether telemetry is enabled (default: true).

public bool Enabled { get; set; }

Property Value

bool

Remarks

For Beginners: Set to true to collect performance metrics, false to disable telemetry. Recommended to keep enabled for production monitoring.

ExportEndpoint

Gets or sets the telemetry export endpoint URL (optional).

public string? ExportEndpoint { get; set; }

Property Value

string

Remarks

For Beginners: URL to send telemetry data for centralized monitoring. Leave null to only store locally. Set to your monitoring system's endpoint.

FlushIntervalSeconds

Gets or sets the telemetry flush interval in seconds (default: 60).

public int FlushIntervalSeconds { get; set; }

Property Value

int

Remarks

For Beginners: How often to send telemetry data to the export endpoint. Default is every minute. Lower values = more real-time, higher values = less network traffic.

SamplingRate

Gets or sets the sampling rate for telemetry (default: 1.0 = 100%).

public double SamplingRate { get; set; }

Property Value

double

Remarks

For Beginners: What percentage of requests to track (0.0 to 1.0). 1.0 = track everything, 0.1 = track 10% of requests. Lower values reduce overhead.

TrackCacheMetrics

Gets or sets whether to track cache hit/miss rates (default: true).

public bool TrackCacheMetrics { get; set; }

Property Value

bool

Remarks

For Beginners: Tracks how often models are found in cache vs loaded from disk. Helps optimize cache size and eviction policies.

TrackDetailedTiming

Gets or sets whether to track detailed timing breakdowns (default: false).

public bool TrackDetailedTiming { get; set; }

Property Value

bool

Remarks

For Beginners: Tracks pre-processing, inference, and post-processing times separately. More detailed but slight performance overhead. Use for debugging performance issues.

TrackErrors

Gets or sets whether to track errors and exceptions (default: true).

public bool TrackErrors { get; set; }

Property Value

bool

Remarks

For Beginners: Records when predictions fail or errors occur. Critical for debugging and maintaining system health.

TrackLatency

Gets or sets whether to track inference latency (default: true).

public bool TrackLatency { get; set; }

Property Value

bool

Remarks

For Beginners: Tracks how long each prediction takes. Helps identify performance problems and slow requests.

TrackThroughput

Gets or sets whether to track throughput metrics (default: true).

public bool TrackThroughput { get; set; }

Property Value

bool

Remarks

For Beginners: Tracks how many predictions per second your system handles. Helps understand capacity and plan for scaling.

TrackVersionUsage

Gets or sets whether to track model version usage (default: true).

public bool TrackVersionUsage { get; set; }

Property Value

bool

Remarks

For Beginners: Records which model versions are being used. Helps track rollouts and A/B test results.