seqcli

Apps that support user-customizable output should use the Seq.Syntax template language. Templates are compiled into ExpressionTemplate instances and rendered against a LogEvent.

Template Compilation

Compile templates in OnAttached() or a factory constructor using an OrderedNameResolver to provide access to built-in functions and custom app properties.

Custom Encoding

If your output format requires escaping (like HTML), implement TemplateOutputEncoder. Note that the encoder is applied only to interpolated values, not to the literal template text.

Built-in Functions and Name Resolution

You can expose custom functions and properties (like @Host or @App) to templates by implementing a NameResolver that maps these tokens to your app's instances.

// Template compilation
using Seq.Syntax.Templates;
using Seq.Syntax.Templates.Encoding;

var template = new ExpressionTemplate(
    templateText,
    nameResolver: new OrderedNameResolver(new NameResolver[]
    {
        new StaticMemberNameResolver(typeof(MyBuiltInFunctions)),
        myAppNameResolver
    }),
    encoder: isPlainText ? null : new MyHtmlEncoder());

// Template rendering
static string Format(ExpressionTemplate template, LogEvent evt)
{
    var writer = new StringWriter();
    template.Format(evt, writer);
    return writer.ToString();
}

// Custom encoding
class TemplateOutputTelegramHtmlEncoder : TemplateOutputEncoder
{
    public override string Encode(string value)
    {
        return value
            .Replace("&", "&")
            .Replace("<", "&lt;")
            .Replace(">", "&gt;");
    }
}

Seq stores OpenTelemetry metric samples in the series data source (a columnar store). To query metrics correctly and avoid performance issues or inaccurate data, you must follow this strict workflow:

1. Search for metric definitions: Use dedicated tools to find metric names and descriptions.
2. Identify the metric kind: Determine if the metric is a Sum, Gauge, Fixed, or Exponential type.
3. List dimensions: Use seqcli metrics dimensions to find which dimensions apply to the metric.
4. List dimension values (Optional): Use seqcli metrics dimensionvalues to find specific values (e.g., service names) within those dimensions.
5. Run aggregate queries: Execute queries against the series data source using the discovered names, kinds, and dimensions.

Critical Constraints

• NEVER query the series data source directly for metric definitions or available dimensions.
• NEVER use distinct to list dimension values; use the dedicated dimension value tools instead.
• NEVER evaluate projection queries against the series data source, as this impacts Seq performance.
• ALWAYS include where Has(<metric name>) in aggregate queries to avoid row count skew.

All events in Seq follow a unified data model. Spans are specifically identified by the presence of the @Start property.

Built-in Properties

seqcli Usage Tip

When using seqcli:

• Use --native to show results in a round-trippable, agent-friendly syntax.
• Use --json for programmatic use, but note that built-in property names are abbreviated in JSON mode.

The Roastery project is a simulation of a coffee bean ordering web application. It models a full HTTP stack using an in-memory data structure that flows from a client through middleware and routing to server-side controllers.

The request flow is: HttpClient $\rightarrow$ Middleware $\rightarrow$ Router $\rightarrow$ Controller.

Key components include:

• HttpClient: A thin wrapper and factory for HttpRequests that acts as a mock web browser.
• HttpServer: The base interface for the middleware and routing layer. It defines the InvokeAsync method.
• Database: A thread-safe in-memory dictionary managing entities like Product and Order.
• Actors: Concurrent entities (e.g., Customer, WarehouseStaff) that drive the simulation by sending requests via the HttpClient.

// The core interface for the HTTP pipeline
abstract class HttpServer
{
    public abstract Task<HttpResponse> InvokeAsync(HttpRequest request);
}