refactor(api): Add more components to healthcheck – follow-ups (#3337)

## What ❔

Various minor follow-ups after
#3193:

- Rework app-level health details.
- Fix `execution_time` unit of measurement for the database health check
details.
- Rework the database health check: do not hold a DB connection all the
time; make it reactive.

## Why ❔

Makes the dependency graph lighter; simplifies maintenance.

## Checklist

- [x] PR title corresponds to the body of PR (we generate changelog
entries from PRs).
- [x] Documentation comments have been added / updated.
- [x] Code has been formatted via `zkstack dev fmt` and `zkstack dev
lint`.

Author: slowli

Cargo.lock

@@ -79,7 +79,6 @@ members = [
   # Test infrastructure
   "core/tests/loadnext",
   "core/tests/vm-benchmark",
-  "core/lib/bin_metadata",
 ]
 resolver = "2"
 

Cargo.toml

@@ -29,7 +29,6 @@ zksync_health_check.workspace = true
 zksync_web3_decl.workspace = true
 zksync_types.workspace = true
 zksync_block_reverter.workspace = true
-zksync_shared_metrics.workspace = true
 zksync_node_genesis.workspace = true
 zksync_node_fee_model.workspace = true
 zksync_node_db_pruner.workspace = true

core/bin/external_node/Cargo.toml

@@ -5,7 +5,6 @@ use zksync_node_framework::{
     implementations::resources::pools::{MasterPool, PoolResource},
     FromContext, IntoContext, StopReceiver, Task, TaskId, WiringError, WiringLayer,
 };
-use zksync_shared_metrics::{GIT_METRICS, RUST_METRICS};
 use zksync_types::{L1ChainId, L2ChainId, SLChainId};
 
 use super::EN_METRICS;
@@ -39,8 +38,6 @@ impl WiringLayer for ExternalNodeMetricsLayer {
     }
 
     async fn wire(self, input: Self::Input) -> Result<Self::Output, WiringError> {
-        RUST_METRICS.initialize();
-        GIT_METRICS.initialize();
         EN_METRICS.observe_config(
             self.l1_chain_id,
             self.sl_chain_id,

core/bin/external_node/src/metrics/framework.rs

@@ -1,6 +1,6 @@
 use std::{collections::HashMap, time::Duration};
 
-use chrono::DateTime;
+use chrono::{DateTime, Utc};
 use serde::{Deserialize, Serialize};
 use zksync_db_connection::{connection::Connection, error::DalResult, instrument::InstrumentExt};
 
@@ -14,11 +14,11 @@ pub(crate) struct TableSize {
     pub total_size: u64,
 }
 
-#[derive(Debug, Serialize, Deserialize)]
+#[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct DatabaseMigration {
     pub version: i64,
     pub description: String,
-    pub installed_on: DateTime<chrono::Utc>,
+    pub installed_on: DateTime<Utc>,
     pub success: bool,
     pub checksum: String,
     pub execution_time: Duration,
@@ -118,7 +118,7 @@ impl SystemDal<'_, '_> {
             installed_on: row.installed_on,
             success: row.success,
             checksum: hex::encode(row.checksum),
-            execution_time: Duration::from_millis(u64::try_from(row.execution_time).unwrap_or(0)),
+            execution_time: Duration::from_nanos(u64::try_from(row.execution_time).unwrap_or(0)),
         })
     }
 }

core/lib/bin_metadata/Cargo.toml

@@ -20,7 +20,6 @@ serde_json.workspace = true
 thiserror.workspace = true
 tokio = { workspace = true, features = ["sync", "time"] }
 tracing.workspace = true
-zksync_bin_metadata.workspace = true
 
 [dev-dependencies]
 assert_matches.workspace = true

core/lib/dal/src/system_dal.rs

@@ -11,12 +11,9 @@ pub use async_trait::async_trait;
 use futures::future;
 use serde::Serialize;
 use tokio::sync::watch;
-use zksync_bin_metadata::BIN_METADATA;
 
-use self::metrics::{CheckResult, METRICS};
-use crate::metrics::AppHealthCheckConfig;
+use crate::metrics::{AppHealthCheckConfig, CheckResult, METRICS};
 
-mod binary;
 mod metrics;
 
 #[cfg(test)]
@@ -114,6 +111,8 @@ pub struct AppHealthCheck {
 
 #[derive(Debug, Clone)]
 struct AppHealthCheckInner {
+    /// Application-level health details.
+    app_details: Option<serde_json::Value>,
     components: Vec<Arc<dyn CheckHealth>>,
     slow_time_limit: Duration,
     hard_time_limit: Duration,
@@ -136,6 +135,7 @@ impl AppHealthCheck {
 
         let inner = AppHealthCheckInner {
             components: Vec::default(),
+            app_details: None,
             slow_time_limit,
             hard_time_limit,
         };
@@ -181,6 +181,13 @@ impl AppHealthCheck {
         }
     }
 
+    /// Sets app-level health details. They can include build info etc.
+    pub fn set_details(&self, details: impl Serialize) {
+        let details = serde_json::to_value(details).expect("failed serializing app details");
+        let mut inner = self.inner.lock().expect("`AppHealthCheck` is poisoned");
+        inner.app_details = Some(details);
+    }
+
     /// Inserts health check for a component.
     ///
     /// # Errors
@@ -220,6 +227,7 @@ impl AppHealthCheck {
         // Clone `inner` so that we don't hold a lock for them across a wait point.
         let AppHealthCheckInner {
             components,
+            app_details,
             slow_time_limit,
             hard_time_limit,
         } = self
@@ -238,7 +246,8 @@ impl AppHealthCheck {
             .map(|health| health.status)
             .max_by_key(|status| status.priority_for_aggregation())
             .unwrap_or(HealthStatus::Ready);
-        let inner = Health::with_details(aggregated_status.into(), BIN_METADATA);
+        let mut inner = Health::from(aggregated_status);
+        inner.details = app_details.clone();
 
         let health = AppHealth { inner, components };
         if !health.inner.status.is_healthy() {

core/lib/health_check/Cargo.toml

@@ -82,6 +82,7 @@ async fn aggregating_health_checks() {
     let (first_check, first_updater) = ReactiveHealthCheck::new("first");
     let (second_check, second_updater) = ReactiveHealthCheck::new("second");
     let inner = AppHealthCheckInner {
+        app_details: None,
         components: vec![Arc::new(first_check), Arc::new(second_check)],
         slow_time_limit: AppHealthCheck::DEFAULT_SLOW_TIME_LIMIT,
         hard_time_limit: AppHealthCheck::DEFAULT_HARD_TIME_LIMIT,