#metrics #open-telemetry #atomic #observable #sync #influxive #obs

influxive-otel-atomic-obs

Opentelemetry observable metric implementations based on std::sync::atomic types

8 releases

0.0.4-alpha.1 Feb 19, 2025
0.0.3-alpha.1 Jun 5, 2024
0.0.2-alpha.1 Oct 16, 2023
0.0.1-alpha.11 Aug 4, 2023
0.0.1-alpha.9 Jul 31, 2023

#616 in Development tools

Download history 518/week @ 2024-11-17 595/week @ 2024-11-24 430/week @ 2024-12-01 486/week @ 2024-12-08 527/week @ 2024-12-15 154/week @ 2024-12-22 193/week @ 2024-12-29 639/week @ 2025-01-05 684/week @ 2025-01-12 517/week @ 2025-01-19 541/week @ 2025-01-26 859/week @ 2025-02-02 754/week @ 2025-02-09 706/week @ 2025-02-16 601/week @ 2025-02-23 673/week @ 2025-03-02

2,887 downloads per month
Used in 20 crates (2 directly)

MIT/Apache

26KB
594 lines

Opentelemetry observable metric implementations based on std::sync::atomic types. Opentelemetry has a concept of "observable" metrics that are not reported as they are updated, but rather, when an update happens, they are polled. For ease-of-use in code, it is often desirable to have these metrics be backed by std::sync::atomic types, so that they can be easily updated throughout the code, and fetched whenever a metric reporting poll occurs. This crate provides the [MeterExt] trait and associated types to make it easy to use std::sync::atomic backed metrics with opentelemetry.

Example

use influxive_otel_atomic_obs::MeterExt;

let (my_metric, _) = opentelemetry_api::global::meter("my_meter")
    .u64_observable_gauge_atomic("my_metric", 0)
    .init();

my_metric.set(66); // probably will not be reported
my_metric.set(99); // probably will not be reported
my_metric.set(42); // will be reported next time reporting runs

Dependencies

~2–3MB
~54K SLoC