Struct MemMetricsBackend 

pub struct MemMetricsBackend { /* private fields */ }

Non-durable in-memory metrics storage.

Counter and gauge writes are appended as points and can be queried immediately. This backend is the default spectra backend and is useful for local development and tests.

Examples

Facade wiring through Spectra::builder() (requires the spectra crate):

use std::sync::Arc;
use spectra::{MemEventsBackend, MemMetricsBackend, Spectra};

let spectra = Spectra::builder()
    .metrics_backend(Arc::new(MemMetricsBackend::new()))
    .events_backend(Arc::new(MemEventsBackend::new()))
    .embedded()
    .build()?;

Direct backend usage:

use chrono::{Duration, Utc};
use serde_json::json;
use spectra_backend_mem::MemMetricsBackend;
use spectra_core::{MetricsQueryRange, MetricsStorageBackend};

let backend = MemMetricsBackend::new();
let now = Utc::now();
backend.record_counter("cache_hits", &json!({"region": "us"}), 1, now).await?;

let points = backend.query_range(MetricsQueryRange {
    metric_name: "cache_hits".into(),
    start: now - Duration::seconds(1),
    end: now + Duration::seconds(1),
    label_matchers: vec![],
}).await?;
assert_eq!(points.len(), 1);

Implementations

impl MemMetricsBackend

pub fn new() -> MemMetricsBackend

Create an empty in-memory metrics backend.

Writes are process-local and discarded when the process exits. Inject the result into Spectra::builder().metrics_backend(...) for a full runtime, or call storage trait methods directly in tests.

Examples

use spectra_backend_mem::MemMetricsBackend;
use spectra_core::{MetricsStorageBackend, StorageEngineType};

let backend = MemMetricsBackend::new();
assert_eq!(backend.engine_type(), StorageEngineType::Mem);

Trait Implementations

impl Default for MemMetricsBackend

fn default() -> MemMetricsBackend

Returns the “default value” for a type.

impl MetricsStorageBackend for MemMetricsBackend

fn engine_type(&self) -> StorageEngineType

Returns the engine type for this backend.

fn record_counter<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, name: &'life1 str, labels: &'life2 Value, delta: i64, ts: DateTime<Utc>, ) -> Pin<Box<dyn Future<Output = Result<(), Error>> + Send + 'async_trait>>

where 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, MemMetricsBackend: 'async_trait,

Records a counter increment.

fn record_gauge<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, name: &'life1 str, labels: &'life2 Value, value: f64, ts: DateTime<Utc>, ) -> Pin<Box<dyn Future<Output = Result<(), Error>> + Send + 'async_trait>>

where 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait, MemMetricsBackend: 'async_trait,

Records a gauge sample.

fn query_range<'life0, 'async_trait>( &'life0 self, query: MetricsQueryRange, ) -> Pin<Box<dyn Future<Output = Result<Vec<MetricPoint>, Error>> + Send + 'async_trait>>

where 'life0: 'async_trait, MemMetricsBackend: 'async_trait,

Queries a time range of metric points.

fn record_metrics_batch<'life0, 'life1, 'async_trait>( &'life0 self, rows: &'life1 [MetricWriteRow], ) -> Pin<Box<dyn Future<Output = Result<(), Error>> + Send + 'async_trait>>

where 'life0: 'async_trait, 'life1: 'async_trait, Self: 'async_trait,

Batch write metrics (default: per-row dispatch).