Marco Napetti


Thank you very much for your support. I've tried every version of hyper-native-tls released so far and the problem doesn't change, I'll look deeper into it asap, in the meanwhile...

I thought about that only now, probably the best solution is to use `Metric::u64_value_recorder` and put always 1 as value, letting the consumer to sum the values in the time...

I was wrong, the sum is still monotonically growing... I need exact values to get what I want, so I'm back to a global hashmap to be rest at every...

I'm sending metrics via otlp to an external service (newrelic). At the moment I solved with a global hashmap and a value observer inside a batch observer. Using a sum_value,...

I init metrics with a methos like that: ```rust fn init_meter(apikey: &str) -> metrics::Result { let mut map = tonic::metadata::MetadataMap::with_capacity(1); map.insert( "api-key", apikey.parse().unwrap(), ); opentelemetry_otlp::new_pipeline() .metrics(spawn, delayed_interval) .with_exporter( opentelemetry_otlp::new_exporter() .tonic()...

To give you a visual example, this is a u64_counter metric, I'm calling the single endpoints 20 times a minute, so I expect it to be a flatline, not to...

I've made another test to clarify which element is failing, and it seems to be opentelemetry-rust. I've attached a prometheus instance to the same collector that sends data to newrelic,...

I've been able to replicate the problem with a "minimal" example: ```rust use futures_util::{Stream, StreamExt, Future}; use hyper::{ body, service::{make_service_fn, service_fn}, Body, Request, Response, Server, }; use once_cell::sync::Lazy; use opentelemetry::{...

Is it possible to output an error log when this happens? It would be very helpful for troubleshooting. Can I help with patch development? Or just with the tests? Il...