refactor(e2e/kms): replace fixed startup sleeps with KMS readiness probe

Replace 33 hard-coded sleep(3s) / sleep(2s) startup waits in KMS e2e tests
with an active readiness probe (wait_for_kms_ready) that polls the KMS
status endpoint with exponential backoff (200ms→1s, 5s budget).

This cuts per-test startup latency from a fixed 3s to ~200-500ms while
remaining robust against slow CI machines.

Non-startup sleeps (ILM polling loops, fault-recovery detection delays,
test-runner inter-test pauses) are left untouched.

crates/e2e_test/src/kms/bucket_default_encryption_test.rs

@@ -37,7 +37,7 @@ async fn test_bucket_default_sse_s3_put_object() -> Result<(), Box<dyn std::erro
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -159,7 +159,7 @@ async fn test_bucket_default_sse_kms_put_object() -> Result<(), Box<dyn std::err
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -278,7 +278,7 @@ async fn test_bucket_default_sse_kms_multipart_crc32() -> Result<(), Box<dyn std
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -475,7 +475,7 @@ async fn test_explicit_encryption_overrides_bucket_default() -> Result<(), Box<d
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -570,7 +570,7 @@ async fn test_sse_kms_without_key_id_populates_default() -> Result<(), Box<dyn s
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;

crates/e2e_test/src/kms/common.rs

@@ -40,7 +40,7 @@ use std::time::Duration;
 use tokio::fs;
 use tokio::net::TcpStream;
 use tokio::time::sleep;
-use tracing::{debug, error, info};
+use tracing::{debug, error, info, warn};
 
 // KMS-specific constants
 pub const TEST_BUCKET: &str = "kms-test-bucket";
@@ -177,6 +177,49 @@ pub async fn get_kms_status(
     Ok(status)
 }
 
+/// Poll the KMS status endpoint until the backend reports ready or the timeout
+/// expires.  Replaces hard-coded `sleep(Duration::from_secs(3))` startup waits
+/// with an active readiness probe so tests start as soon as KMS is usable
+/// (typically < 1 s) instead of always waiting the full 3 s.
+///
+/// Uses exponential back-off starting at 200 ms (doubling each attempt, capped
+/// at 1 s) up to a total wall-clock budget of 5 s.
+pub async fn wait_for_kms_ready(
+    base_url: &str,
+    access_key: &str,
+    secret_key: &str,
+) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+    let total_deadline = Duration::from_secs(5);
+    let start = tokio::time::Instant::now();
+    let mut backoff = Duration::from_millis(200);
+    let max_backoff = Duration::from_secs(1);
+    let mut first_attempt = true;
+
+    loop {
+        if !first_attempt {
+            if start.elapsed() >= total_deadline {
+                return Err("KMS failed to become ready within 5 seconds".into());
+            }
+            sleep(backoff).await;
+            backoff = (backoff * 2).min(max_backoff);
+        }
+        first_attempt = false;
+
+        match get_kms_status(base_url, access_key, secret_key).await {
+            Ok(status) => {
+                info!("KMS is ready (status: {})", status);
+                return Ok(());
+            }
+            Err(e) => {
+                if start.elapsed() >= total_deadline {
+                    return Err(format!("KMS did not become ready within 5 s: last error: {e}").into());
+                }
+                warn!(error = %e, elapsed_ms = start.elapsed().as_millis() as u64, "KMS not ready yet, retrying…");
+            }
+        }
+    }
+}
+
 /// Create a default KMS key for testing and return the created key ID
 pub async fn create_default_key(
     base_url: &str,
@@ -861,6 +904,13 @@ impl LocalKMSTestEnvironment {
         Ok(default_key_id.to_string())
     }
 
+    /// Poll the KMS status endpoint until the backend reports ready.
+    ///
+    /// Prefer this over a fixed `sleep` after calling `start_rustfs_for_local_kms`.
+    pub async fn wait_for_kms_ready(&self) -> Result<(), Box<dyn std::error::Error + Send + Sync>> {
+        wait_for_kms_ready(&self.base_env.url, &self.base_env.access_key, &self.base_env.secret_key).await
+    }
+
     /// Configure Local KMS backend with a predefined default key
     pub async fn configure_local_kms(&self) -> Result<String, Box<dyn std::error::Error + Send + Sync>> {
         // Use a fixed, predictable default key ID

crates/e2e_test/src/kms/copy_object_self_copy_sse_test.rs

@@ -61,7 +61,7 @@ async fn test_metadata_replace_self_copy_of_sse_object_stays_decryptable() {
         )
         .await
         .expect("failed to start RustFS with local KMS");
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let client = kms_env.base_env.create_s3_client();
     // Deliberately an UNVERSIONED bucket: that is the branch where the store layer can service
@@ -160,7 +160,7 @@ async fn test_metadata_replace_self_copy_dropping_sse_rewrites_plaintext() {
         )
         .await
         .expect("failed to start RustFS with local KMS");
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let client = kms_env.base_env.create_s3_client();
     // Unversioned, and deliberately WITHOUT a bucket default-encryption rule, so the copy below
@@ -256,7 +256,7 @@ async fn test_metadata_replace_self_copy_under_bucket_default_sse_stays_decrypta
         )
         .await
         .expect("failed to start RustFS with local KMS");
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let client = kms_env.base_env.create_s3_client();
     let bucket = "copy-object-self-copy-bucket-default-sse-test";

crates/e2e_test/src/kms/copy_object_version_restore_sse_test.rs

@@ -56,7 +56,7 @@ async fn test_self_copy_of_historical_sse_s3_version_is_readable() {
         )
         .await
         .expect("failed to start RustFS with local KMS");
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let client = kms_env.base_env.create_s3_client();
     let bucket = "copy-object-version-restore-sse-test";

crates/e2e_test/src/kms/encryption_metadata_test.rs

@@ -87,7 +87,7 @@ async fn test_head_reports_managed_metadata_for_sse_s3() -> Result<(), Box<dyn s
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -147,7 +147,7 @@ async fn test_head_reports_managed_metadata_for_sse_kms_and_copy() -> Result<(),
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -250,7 +250,7 @@ async fn test_multipart_upload_writes_encrypted_data() -> Result<(), Box<dyn std
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;

crates/e2e_test/src/kms/kms_comprehensive_test.rs

@@ -24,7 +24,6 @@ use super::common::{
     test_sse_kms_encryption, test_sse_s3_encryption,
 };
 use crate::common::{TEST_BUCKET, init_logging};
-use tokio::time::{Duration, sleep};
 use tracing::info;
 
 /// Comprehensive test: Full KMS workflow with all encryption types
@@ -35,7 +34,7 @@ async fn test_comprehensive_kms_full_workflow() -> Result<(), Box<dyn std::error
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -103,7 +102,7 @@ async fn test_comprehensive_stress_test() -> Result<(), Box<dyn std::error::Erro
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -137,7 +136,7 @@ async fn test_comprehensive_key_isolation() -> Result<(), Box<dyn std::error::Er
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -208,7 +207,7 @@ async fn test_comprehensive_concurrent_operations() -> Result<(), Box<dyn std::e
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -253,7 +252,7 @@ async fn test_comprehensive_performance_benchmark() -> Result<(), Box<dyn std::e
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;

crates/e2e_test/src/kms/kms_edge_cases_test.rs

@@ -44,7 +44,7 @@ async fn test_kms_zero_byte_file_encryption() -> Result<(), Box<dyn std::error::
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -117,7 +117,7 @@ async fn test_kms_single_byte_file_encryption() -> Result<(), Box<dyn std::error
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -209,7 +209,7 @@ async fn test_kms_multipart_boundary_conditions() -> Result<(), Box<dyn std::err
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -284,7 +284,7 @@ async fn test_kms_invalid_key_scenarios() -> Result<(), Box<dyn std::error::Erro
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -371,7 +371,7 @@ async fn test_kms_concurrent_encryption() -> Result<(), Box<dyn std::error::Erro
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = Arc::new(kms_env.base_env.create_s3_client());
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -478,7 +478,7 @@ async fn test_kms_key_validation_security() -> Result<(), Box<dyn std::error::Er
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;

crates/e2e_test/src/kms/kms_fault_recovery_test.rs

@@ -37,7 +37,7 @@ async fn test_kms_key_directory_unavailable() -> Result<(), Box<dyn std::error::
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -127,7 +127,7 @@ async fn test_kms_corrupted_key_files() -> Result<(), Box<dyn std::error::Error
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -218,7 +218,7 @@ async fn test_kms_multipart_upload_interruption() -> Result<(), Box<dyn std::err
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -401,7 +401,7 @@ async fn test_kms_resource_constraints() -> Result<(), Box<dyn std::error::Error
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;

crates/e2e_test/src/kms/kms_local_test.rs

@@ -46,7 +46,7 @@ async fn test_local_kms_end_to_end() -> Result<(), Box<dyn std::error::Error + S
         .expect("Failed to start RustFS with Local KMS");
 
     // Wait a moment for RustFS to fully start up and initialize KMS
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     info!("RustFS started with KMS auto-configuration, default_key_id: {}", default_key_id);
 
@@ -127,7 +127,7 @@ async fn test_local_kms_key_isolation() {
         .expect("Failed to start RustFS with Local KMS");
 
     // Wait a moment for RustFS to fully start up and initialize KMS
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     info!("RustFS started with KMS auto-configuration, default_key_id: {}", default_key_id);
 
@@ -227,7 +227,7 @@ async fn test_local_kms_large_file() {
         .expect("Failed to start RustFS with Local KMS");
 
     // Wait a moment for RustFS to fully start up and initialize KMS
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     info!("RustFS started with KMS auto-configuration, default_key_id: {}", default_key_id);
 
@@ -309,7 +309,7 @@ async fn test_local_kms_multipart_upload() {
         .expect("Failed to start RustFS with Local KMS");
 
     // Wait for KMS initialization
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     info!("RustFS started with KMS auto-configuration, default_key_id: {}", default_key_id);
 

crates/e2e_test/src/kms/kms_vault_test.rs

@@ -20,7 +20,6 @@
 
 use crate::common::{TEST_BUCKET, init_logging};
 use serial_test::serial;
-use tokio::time::{Duration, sleep};
 use tracing::{error, info};
 
 use super::common::{
@@ -46,8 +45,13 @@ impl VaultKmsTestContext {
 
         start_kms(&env.base_env.url, &env.base_env.access_key, &env.base_env.secret_key).await?;
 
-        // Allow Vault to finish initialising token auth and transit engine.
-        sleep(Duration::from_secs(2)).await;
+        // Wait for KMS to finish initialising.
+        super::common::wait_for_kms_ready(
+            &env.base_env.url,
+            &env.base_env.access_key,
+            &env.base_env.secret_key,
+        )
+        .await?;
 
         Ok(Self { env })
     }

crates/e2e_test/src/kms/multipart_encryption_test.rs

@@ -33,7 +33,7 @@ async fn test_step1_basic_single_file_encryption() -> Result<(), Box<dyn std::er
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -89,7 +89,7 @@ async fn test_step2_basic_multipart_upload_without_encryption() -> Result<(), Bo
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -187,7 +187,7 @@ async fn test_step3_multipart_upload_with_sse_s3() -> Result<(), Box<dyn std::er
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -310,7 +310,7 @@ async fn test_step4_large_multipart_upload_with_encryption() -> Result<(), Box<d
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;
@@ -435,7 +435,7 @@ async fn test_step5_all_encryption_types_multipart() -> Result<(), Box<dyn std::
 
     let mut kms_env = LocalKMSTestEnvironment::new().await?;
     let _default_key_id = kms_env.start_rustfs_for_local_kms().await?;
-    tokio::time::sleep(tokio::time::Duration::from_secs(3)).await;
+    kms_env.wait_for_kms_ready().await?;
 
     let s3_client = kms_env.base_env.create_s3_client();
     kms_env.base_env.create_test_bucket(TEST_BUCKET).await?;