Add build_info_metric with git_version and build_tag for compute_ctl

2026-05-14 03:30:36 +00:00 · 2023-11-28 17:23:24 +00:00
6 changed files with 39 additions and 2 deletions
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -1133,7 +1133,9 @@ dependencies = [
 "compute_api",
 "flate2",
 "futures",
+ "git-version",
 "hyper",
+ "metrics",
 "notify",
 "num_cpus",
 "opentelemetry",
--- a/compute_tools/Cargo.toml
+++ b/compute_tools/Cargo.toml
@@ -12,8 +12,10 @@ cfg-if.workspace = true
 clap.workspace = true
 flate2.workspace = true
 futures.workspace = true
+git-version.workspace = true
 hyper = { workspace = true, features = ["full"] }
 notify.workspace = true
+metrics.workspace = true
 num_cpus.workspace = true
 opentelemetry.workspace = true
 postgres.workspace = true
--- a/compute_tools/src/bin/compute_ctl.rs
+++ b/compute_tools/src/bin/compute_ctl.rs
@@ -57,18 +57,27 @@ use compute_tools::logger::*;
 use compute_tools::monitor::launch_monitor;
 use compute_tools::params::*;
 use compute_tools::spec::*;
+use metrics::set_build_info_metric;
+use utils::{project_build_tag, project_git_version};

 // this is an arbitrary build tag. Fine as a default / for testing purposes
 // in-case of not-set environment var
 const BUILD_TAG_DEFAULT: &str = "latest";

+project_git_version!(GIT_VERSION);
+project_build_tag!(BUILD_TAG);
+
 fn main() -> Result<()> {
    init_tracing_and_logging(DEFAULT_LOG_LEVEL)?;

    let build_tag = option_env!("BUILD_TAG")
        .unwrap_or(BUILD_TAG_DEFAULT)
        .to_string();
-    info!("build_tag: {build_tag}");
+
+    info!("Version: {GIT_VERSION}");
+    info!("Build_tag: {BUILD_TAG}");
+
+    set_build_info_metric(GIT_VERSION, BUILD_TAG);

    let matches = cli().get_matches();
    let pgbin_default = String::from("postgres");
--- a/compute_tools/src/http/api.rs
+++ b/compute_tools/src/http/api.rs
@@ -11,7 +11,8 @@ use compute_api::responses::{ComputeStatus, ComputeStatusResponse, GenericAPIErr

 use anyhow::Result;
 use hyper::service::{make_service_fn, service_fn};
-use hyper::{Body, Method, Request, Response, Server, StatusCode};
+use hyper::{header::CONTENT_TYPE, Body, Method, Request, Response, Server, StatusCode};
+use metrics::{Encoder, TextEncoder};
 use num_cpus;
 use serde_json;
 use tokio::task;
@@ -51,6 +52,20 @@ async fn routes(req: Request<Body>, compute: &Arc<ComputeNode>) -> Response<Body
            Response::new(Body::from(serde_json::to_string(&status_response).unwrap()))
        }

+        // prometheus metrics
+        (&Method::GET, "/metrics") => {
+            let mut buffer = vec![];
+            let metrics = metrics::gather();
+            let encoder = TextEncoder::new();
+            encoder.encode(&metrics, &mut buffer).unwrap();
+
+            Response::builder()
+                .status(StatusCode::OK)
+                .header(CONTENT_TYPE, encoder.format_type())
+                .body(Body::from(buffer))
+                .unwrap()
+        }
+
        // Startup metrics in JSON format. Keep /metrics reserved for a possible
        // future use for Prometheus metrics format.
        (&Method::GET, "/metrics.json") => {
--- a/test_runner/fixtures/neon_fixtures.py
+++ b/test_runner/fixtures/neon_fixtures.py
@@ -2640,6 +2640,11 @@ class Endpoint(PgProtocol):

        return self

+    def get_metrics_str(self) -> str:
+        request_result = requests.get(f"http://localhost:{self.http_port}/metrics")
+        request_result.raise_for_status()
+        return request_result.text
+
    def __enter__(self) -> "Endpoint":
        return self

--- a/test_runner/regress/test_build_info_metric.py
+++ b/test_runner/regress/test_build_info_metric.py
@@ -6,11 +6,15 @@ def test_build_info_metric(neon_env_builder: NeonEnvBuilder, link_proxy: NeonPro
    neon_env_builder.num_safekeepers = 1
    env = neon_env_builder.init_start()

+    env.neon_cli.create_branch("test_build_info_metric")
+    endpoint = env.endpoints.create_start("test_build_info_metric")
+
    parsed_metrics = {}

    parsed_metrics["pageserver"] = parse_metrics(env.pageserver.http_client().get_metrics_str())
    parsed_metrics["safekeeper"] = parse_metrics(env.safekeepers[0].http_client().get_metrics_str())
    parsed_metrics["proxy"] = parse_metrics(link_proxy.get_metrics())
+    parsed_metrics["compute_ctl"] = parse_metrics(endpoint.get_metrics_str())

    for _component, metrics in parsed_metrics.items():
        sample = metrics.query_one("libmetrics_build_info")