Crate enya_client

Expand description

Metrics client abstraction supporting multiple backends.

This crate provides a unified interface for querying metrics from different backends (Prometheus, Enya, etc.) using PromQL as the query language.

§Architecture

The MetricsClient trait defines a promise-based async interface that all backends implement. Methods return Promise objects that can be polled each frame in immediate mode GUIs like egui. HTTP requests are handled by reqwest which works on both native (with tokio) and WASM (with browser fetch).

§Example

use enya_client::{MetricsClient, QueryRequest};
use enya_client::prometheus::PrometheusClient;

// Create a client for your backend
let client = PrometheusClient::new("http://localhost:9090");

// Fire off a query - returns a promise
let request = QueryRequest::new("cpu_usage", r#"sum(cpu_usage{env="prod"}) by (host)"#);
let promise = client.query(request, &ctx);

// In your update loop, poll for results
if let Some(result) = promise.ready() {
    match result {
        Ok(response) => { /* update visualization */ }
        Err(e) => { /* show error */ }
    }
}

Re-exports§

pub use demo::DemoMetricsClient;
pub use error::ClientError;
pub use promise::promise_channel;
pub use request::QueryRequest;
pub use types::MetricsBucket;
pub use types::MetricsGroup;
pub use types::QueryResponse;
pub use types::ResultType;
pub use types::Timestamp;
pub use prometheus::response::MetricLabels;
pub use logs::DemoLogsClient;
pub use logs::LogEntry;
pub use logs::LogLevel;
pub use logs::LogsClient;
pub use logs::LogsQuery;
pub use logs::LogsResponse;
pub use logs::LogsResult;
pub use logs::LokiClient;
pub use logs::QueryDirection;
pub use logs::StreamsResult;

Modules§

demo: Demo metrics client for offline/showcase mode.
error: Error types for the metrics client.
logs: Logs client abstraction supporting multiple backends.
otlp: OpenTelemetry Protocol (OTLP) support.
prometheus: Prometheus backend implementation.
promise: Promise-based async helpers for HTTP requests.
request: Query request types.
tracing: Tracing client abstraction for distributed trace backends.
types: Shared API types for metrics query responses.

Structs§

BackendInfo: Backend health/version information from a health check.
HealthCheckManager: Manages in-flight health check requests using promises.
LabelsManager: Manages in-flight label/metadata fetches using promises.
LogsQueryManager: Manages multiple in-flight log queries in parallel using promises.
MetricLabelsManager: Manages in-flight per-metric label fetches using promises.
Promise: A promise that waits for the reception of a single value, presumably from some async task.
QueryManager: Manages multiple in-flight queries in parallel using promises.

Constants§

DEFAULT_QUERY_TIMEOUT_SECS: Default query timeout in seconds. If a query doesn’t complete within this time, it will be cancelled with a timeout error.

Traits§

MetricsClient: Metrics client trait - promise-based async interface.

Functions§

normalize_url: Normalize a base URL: ensure it has an http:// scheme and strip trailing slashes.
now_unix_secs: Get the current Unix timestamp in seconds. Works on both native and WASM platforms.
url_encode: Simple URL encoding for query parameters.

Type Aliases§

HealthCheckResult: Result type for health check operations.
LabelsResult: Result type for label list operations.
MetricLabelsResult: Result type for metric series label operations.
QueryResult: Result type for query operations.