Make pageserver_ prefix for common metric names configurable (#681)

Cargo.lock

@@ -2578,6 +2578,7 @@ version = "0.1.0"
 dependencies = [
  "lazy_static",
  "libc",
+ "once_cell",
  "prometheus",
 ]
 

pageserver/src/bin/pageserver.rs

@@ -240,6 +240,7 @@ impl CfgFileParams {
 }
 
 fn main() -> Result<()> {
+    zenith_metrics::set_common_metrics_prefix("pageserver");
     let arg_matches = App::new("Zenith page server")
         .about("Materializes WAL stream to pages and serves them to the postgres")
         .arg(

test_runner/fixtures/benchmark_fixture.py

@@ -136,7 +136,8 @@ class ZenithBenchmarker:
         # The metric should be an integer, as it's a number of bytes. But in general
         # all prometheus metrics are floats. So to be pedantic, read it as a float
         # and round to integer.
-        matches = re.search(r'pageserver_disk_io_bytes{io_operation="write"} (\S+)', all_metrics)
+        matches = re.search(r'^pageserver_disk_io_bytes{io_operation="write"} (\S+)$', all_metrics,
+                            re.MULTILINE)
         return int(round(float(matches.group(1))))
 
     @contextmanager

zenith_metrics/Cargo.toml

@@ -7,3 +7,4 @@ edition = "2018"
 prometheus = {version = "0.12", default_features=false} # removes protobuf dependency
 libc = "0.2"
 lazy_static = "1.4"
+once_cell = "1.8.0"

zenith_metrics/src/lib.rs

@@ -3,6 +3,7 @@
 //! Otherwise, we might not see all metrics registered via
 //! a default registry.
 use lazy_static::lazy_static;
+use once_cell::race::OnceBox;
 pub use prometheus::{exponential_buckets, linear_buckets};
 pub use prometheus::{register_histogram, Histogram};
 pub use prometheus::{register_histogram_vec, HistogramVec};
@@ -24,9 +25,29 @@ pub fn gather() -> Vec<prometheus::proto::MetricFamily> {
     prometheus::gather()
 }
 
+static COMMON_METRICS_PREFIX: OnceBox<&str> = OnceBox::new();
+
+/// Sets a prefix which will be used for all common metrics, typically a service
+/// name like 'pageserver'. Should be executed exactly once in the beginning of
+/// any executable which uses common metrics.
+pub fn set_common_metrics_prefix(prefix: &'static str) {
+    COMMON_METRICS_PREFIX.set(prefix.into()).unwrap();
+}
+
+/// Prepends a prefix to a common metric name so they are distinguished between
+/// different services, see https://github.com/zenithdb/zenith/pull/681
+/// A call to set_common_metrics_prefix() is necessary prior to calling this.
+pub fn new_common_metric_name(unprefixed_metric_name: &str) -> String {
+    format!(
+        "{}_{}",
+        COMMON_METRICS_PREFIX.get().unwrap(),
+        unprefixed_metric_name
+    )
+}
+
 lazy_static! {
     static ref DISK_IO_BYTES: IntGaugeVec = register_int_gauge_vec!(
-        "pageserver_disk_io_bytes",
+        new_common_metric_name("disk_io_bytes"),
         "Bytes written and read from disk, grouped by the operation (read|write)",
         &["io_operation"]
     )

zenith_utils/src/http/endpoint.rs

@@ -9,14 +9,14 @@ use routerify::ext::RequestExt;
 use routerify::RequestInfo;
 use routerify::{Middleware, Router, RouterBuilder, RouterService};
 use std::net::TcpListener;
-use zenith_metrics::{register_int_counter, IntCounter};
+use zenith_metrics::{new_common_metric_name, register_int_counter, IntCounter};
 use zenith_metrics::{Encoder, TextEncoder};
 
 use super::error::ApiError;
 
 lazy_static! {
     static ref SERVE_METRICS_COUNT: IntCounter = register_int_counter!(
-        "pageserver_serve_metrics_count",
+        new_common_metric_name("serve_metrics_count"),
         "Number of metric requests made"
     )
     .expect("failed to define a metric");