WIP: opentelemetry-0.18 bump #490

Closed
jpds wants to merge 12 commits from jpds/garage:opentelemetry-0.18 into main
2 changed files with 46 additions and 72 deletions
Showing only changes of commit aea585b504 - Show all commits

View file

@ -3,7 +3,6 @@ use std::collections::HashMap;
use std::io::{Read, Write};
use std::net::{IpAddr, SocketAddr};
use std::path::{Path, PathBuf};
use std::sync::atomic::Ordering;
use std::sync::{Arc, RwLock};
use std::time::{Duration, Instant};
@ -11,6 +10,7 @@ use arc_swap::ArcSwap;
use async_trait::async_trait;
use futures::{join, select};
use futures_util::future::*;
use opentelemetry::{Context, KeyValue};
use serde::{Deserialize, Serialize};
use sodiumoxide::crypto::sign::ed25519;
use tokio::sync::watch;
@ -248,7 +248,12 @@ impl System {
replication_mode: ReplicationMode,
config: &Config,
) -> Result<Arc<Self>, Error> {
let metrics = SystemMetrics::new();
let replication_factor = replication_mode.replication_factor();
metrics
._replication_factor
.observe(&Context::current(), replication_factor as u64, &[]);
let node_key =
gen_node_key(&config.metadata_dir).expect("Unable to read or generate node ID");
@ -281,7 +286,14 @@ impl System {
}
};
let metrics = SystemMetrics::new(replication_factor);
metrics._garage_build_info.observe(
&Context::current(),
1,
&[KeyValue::new(
"version",
garage_util::version::garage_version(),
)],
);
let mut local_status = NodeStatus::initial(replication_factor, &cluster_layout);
local_status.update_disk_usage(&config.metadata_dir, &config.data_dir, &metrics);
@ -892,6 +904,7 @@ impl NodeStatus {
fn update_disk_usage(&mut self, meta_dir: &Path, data_dir: &Path, metrics: &SystemMetrics) {
use systemstat::{Platform, System};
let mounts = System::new().mounts().unwrap_or_default();
let mount_avail = |path: &Path| {
@ -906,24 +919,28 @@ impl NodeStatus {
self.data_disk_avail = mount_avail(data_dir);
if let Some((avail, total)) = self.meta_disk_avail {
metrics
.values
.meta_disk_avail
.store(avail, Ordering::Relaxed);
metrics
.values
.meta_disk_total
.store(total, Ordering::Relaxed);
metrics._disk_avail.observe(
&Context::current(),
avail,
&[KeyValue::new("volume", "meta")],
);
metrics._disk_total.observe(
&Context::current(),
total,
&[KeyValue::new("volume", "meta")],
);
}
if let Some((avail, total)) = self.data_disk_avail {
metrics
.values
.data_disk_avail
.store(avail, Ordering::Relaxed);
metrics
.values
.data_disk_total
.store(total, Ordering::Relaxed);
metrics._disk_avail.observe(
&Context::current(),
avail,
&[KeyValue::new("volume", "data")],
);
metrics._disk_total.observe(
&Context::current(),
total,
&[KeyValue::new("volume", "data")],
);
}
}
}

View file

@ -1,77 +1,34 @@
use std::sync::atomic::{AtomicU64, Ordering};
use std::sync::Arc;
use opentelemetry::{global, metrics::*, KeyValue};
use opentelemetry::{global, metrics::*};
/// TableMetrics reference all counter used for metrics
pub struct SystemMetrics {
pub(crate) _garage_build_info: ValueObserver<u64>,
pub(crate) _replication_factor: ValueObserver<u64>,
pub(crate) _disk_avail: ValueObserver<u64>,
pub(crate) _disk_total: ValueObserver<u64>,
pub(crate) values: Arc<SystemMetricsValues>,
}
#[derive(Default)]
pub struct SystemMetricsValues {
pub(crate) data_disk_total: AtomicU64,
pub(crate) data_disk_avail: AtomicU64,
pub(crate) meta_disk_total: AtomicU64,
pub(crate) meta_disk_avail: AtomicU64,
pub(crate) _garage_build_info: ObservableGauge<u64>,
pub(crate) _replication_factor: ObservableGauge<u64>,
pub(crate) _disk_avail: ObservableGauge<u64>,
pub(crate) _disk_total: ObservableGauge<u64>,
}
impl SystemMetrics {
pub fn new(replication_factor: usize) -> Self {
pub fn new() -> Self {
let meter = global::meter("garage_system");
let values = Arc::new(SystemMetricsValues::default());
let values1 = values.clone();
let values2 = values.clone();
Self {
_garage_build_info: meter
.u64_value_observer("garage_build_info", move |observer| {
observer.observe(
1,
&[KeyValue::new(
"version",
garage_util::version::garage_version(),
)],
)
})
.u64_observable_gauge("garage_build_info")
.with_description("Garage build info")
.init(),
_replication_factor: meter
.u64_value_observer("garage_replication_factor", move |observer| {
observer.observe(replication_factor as u64, &[])
})
.u64_observable_gauge("garage_replication_factor")
.with_description("Garage replication factor setting")
.init(),
_disk_avail: meter
.u64_value_observer("garage_local_disk_avail", move |observer| {
match values1.data_disk_avail.load(Ordering::Relaxed) {
0 => (),
x => observer.observe(x, &[KeyValue::new("volume", "data")]),
};
match values1.meta_disk_avail.load(Ordering::Relaxed) {
0 => (),
x => observer.observe(x, &[KeyValue::new("volume", "metadata")]),
};
})
.u64_observable_gauge("garage_local_disk_avail")
.with_description("Garage available disk space on each node")
.init(),
_disk_total: meter
.u64_value_observer("garage_local_disk_total", move |observer| {
match values2.data_disk_total.load(Ordering::Relaxed) {
0 => (),
x => observer.observe(x, &[KeyValue::new("volume", "data")]),
};
match values2.meta_disk_total.load(Ordering::Relaxed) {
0 => (),
x => observer.observe(x, &[KeyValue::new("volume", "metadata")]),
};
})
.u64_observable_gauge("garage_local_disk_total")
.with_description("Garage total disk space on each node")
.init(),
values,
}
}
}