ci: ensure commits from main branch for whitelisted git dependencies (#7434)

* chore: update proto to include native histogram

* ci: add a CI check to ensure whitelisted dependencies are using their main branch

* chore: add changes to Cargo.toml to trigger CI

* chore: update proto

* test: update test to include histogram

.github/workflows/check-git-deps.yml

@@ -0,0 +1,154 @@
+name: Check Git Dependencies on Main Branch
+
+on:
+    pull_request:
+        branches: [main]
+        paths:
+            - 'Cargo.toml'
+    push:
+        branches: [main]
+        paths:
+            - 'Cargo.toml'
+
+jobs:
+    check-git-deps:
+        runs-on: ubuntu-latest
+
+        steps:
+            - name: Checkout repository
+              uses: actions/checkout@v6
+
+            - name: Check git dependencies
+              env:
+                  WHITELIST_DEPS: "greptime-proto,meter-core,meter-macros"
+              run: |
+                  #!/bin/bash
+                  set -e
+
+                  echo "Checking whitelisted git dependencies..."
+
+                  # Function to check if a commit is on main branch
+                  check_commit_on_main() {
+                      local repo_url="$1"
+                      local commit="$2"
+                      local repo_name=$(basename "$repo_url" .git)
+
+                      echo "Checking $repo_name"
+                      echo "Repo: $repo_url"
+                      echo "Commit: $commit"
+
+                      # Create a temporary directory for cloning
+                      local temp_dir=$(mktemp -d)
+
+                      # Clone the repository
+                      if git clone "$repo_url" "$temp_dir" 2>/dev/null; then
+                          cd "$temp_dir"
+
+                          # Try to determine the main branch name
+                          local main_branch="main"
+                          if ! git rev-parse --verify origin/main >/dev/null 2>&1; then
+                              if git rev-parse --verify origin/master >/dev/null 2>&1; then
+                                  main_branch="master"
+                              else
+                                  # Try to get the default branch
+                                  main_branch=$(git symbolic-ref refs/remotes/origin/HEAD | sed 's@^refs/remotes/origin/@@')
+                              fi
+                          fi
+
+                          echo "Main branch: $main_branch"
+
+                          # Check if commit exists
+                          if git cat-file -e "$commit" 2>/dev/null; then
+                              # Check if commit is on main branch
+                              if git merge-base --is-ancestor "$commit" "origin/$main_branch" 2>/dev/null; then
+                                  echo "PASS: Commit $commit is on $main_branch branch"
+                                  cd - >/dev/null
+                                  rm -rf "$temp_dir"
+                                  return 0
+                              else
+                                  echo "FAIL: Commit $commit is NOT on $main_branch branch"
+
+                                  # Try to find which branch contains this commit
+                                  local branch_name=$(git branch -r --contains "$commit" 2>/dev/null | head -1 | sed 's/^[[:space:]]*origin\///' | sed 's/[[:space:]]*$//')
+                                  if [[ -n "$branch_name" ]]; then
+                                      echo "Found on branch: $branch_name"
+                                  fi
+                                  cd - >/dev/null
+                                  rm -rf "$temp_dir"
+                                  return 1
+                              fi
+                          else
+                              echo "FAIL: Commit $commit not found in repository"
+                              cd - >/dev/null
+                              rm -rf "$temp_dir"
+                              return 1
+                          fi
+                      else
+                          echo "FAIL: Failed to clone $repo_url"
+                          rm -rf "$temp_dir"
+                          return 1
+                      fi
+                  }
+
+                  # Extract whitelisted git dependencies from Cargo.toml
+                  echo "Extracting git dependencies from Cargo.toml..."
+
+                  # Create temporary array to store dependencies
+                  declare -a deps=()
+
+                  # Build awk pattern from whitelist
+                  IFS=',' read -ra WHITELIST <<< "$WHITELIST_DEPS"
+                  awk_pattern=""
+                  for dep in "${WHITELIST[@]}"; do
+                      if [[ -n "$awk_pattern" ]]; then
+                          awk_pattern="$awk_pattern|"
+                      fi
+                      awk_pattern="$awk_pattern$dep"
+                  done
+
+                  # Extract whitelisted dependencies
+                  while IFS= read -r line; do
+                      if [[ -n "$line" ]]; then
+                          deps+=("$line")
+                      fi
+                  done < <(awk -v pattern="$awk_pattern" '
+                  $0 ~ pattern ".*git = \"https:/" {
+                      match($0, /git = "([^"]+)"/, arr)
+                      git_url = arr[1]
+                      if (match($0, /rev = "([^"]+)"/, rev_arr)) {
+                          rev = rev_arr[1]
+                          print git_url " " rev
+                      } else {
+                          # Check next line for rev
+                          getline
+                          if (match($0, /rev = "([^"]+)"/, rev_arr)) {
+                              rev = rev_arr[1]
+                              print git_url " " rev
+                          }
+                      }
+                  }
+                  ' Cargo.toml)
+
+                  echo "Found ${#deps[@]} dependencies to check:"
+                  for dep in "${deps[@]}"; do
+                      echo "  $dep"
+                  done
+
+                  failed=0
+
+                  for dep in "${deps[@]}"; do
+                      read -r repo_url commit <<< "$dep"
+                      if ! check_commit_on_main "$repo_url" "$commit"; then
+                          failed=1
+                      fi
+                  done
+
+                  echo "Check completed."
+
+                  if [[ $failed -eq 1 ]]; then
+                      echo "ERROR: Some git dependencies are not on their main branches!"
+                      echo "Please update the commits to point to main branch commits."
+                      exit 1
+                  else
+                      echo "SUCCESS: All git dependencies are on their main branches!"
+                  fi

Cargo.lock

@@ -5461,7 +5461,7 @@ dependencies = [
 [[package]]
 name = "greptime-proto"
 version = "0.1.0"
-source = "git+https://github.com/GreptimeTeam/greptime-proto.git?rev=0423fa30203187c75e2937a668df1da699c8b96c#0423fa30203187c75e2937a668df1da699c8b96c"
+source = "git+https://github.com/GreptimeTeam/greptime-proto.git?rev=173efe5ec62722089db7c531c0b0d470a072b915#173efe5ec62722089db7c531c0b0d470a072b915"
 dependencies = [
  "prost 0.13.5",
  "prost-types 0.13.5",

Cargo.toml

@@ -150,7 +150,7 @@ etcd-client = { version = "0.16.1", features = [
 fst = "0.4.7"
 futures = "0.3"
 futures-util = "0.3"
-greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "0423fa30203187c75e2937a668df1da699c8b96c" }
+greptime-proto = { git = "https://github.com/GreptimeTeam/greptime-proto.git", rev = "173efe5ec62722089db7c531c0b0d470a072b915" }
 hex = "0.4"
 http = "1"
 humantime = "2.1"

config/datanode.example.toml

@@ -225,6 +225,7 @@ overwrite_entry_start_id = false
 # endpoint = "https://s3.amazonaws.com"
 # region = "us-west-2"
 # enable_virtual_host_style = false
+# disable_ec2_metadata = false
 
 # Example of using Oss as the storage.
 # [storage]

config/standalone.example.toml

@@ -332,6 +332,7 @@ max_running_procedures = 128
 # endpoint = "https://s3.amazonaws.com"
 # region = "us-west-2"
 # enable_virtual_host_style = false
+# disable_ec2_metadata = false
 
 # Example of using Oss as the storage.
 # [storage]

src/cli/src/common/object_store.rs

@@ -255,6 +255,8 @@ wrap_with_clap_prefix! {
         region: Option<String>,
         #[doc = "Enable virtual host style for the object store."]
         enable_virtual_host_style: bool = Default::default(),
+        #[doc = "Disable EC2 metadata service for the object store."]
+        disable_ec2_metadata: bool = Default::default(),
     }
 }
 

src/common/datasource/src/object_store/s3.rs

@@ -27,6 +27,7 @@ const SECRET_ACCESS_KEY: &str = "secret_access_key";
 const SESSION_TOKEN: &str = "session_token";
 const REGION: &str = "region";
 const ENABLE_VIRTUAL_HOST_STYLE: &str = "enable_virtual_host_style";
+const DISABLE_EC2_METADATA: &str = "disable_ec2_metadata";
 
 pub fn is_supported_in_s3(key: &str) -> bool {
     [
@@ -36,6 +37,7 @@ pub fn is_supported_in_s3(key: &str) -> bool {
         SESSION_TOKEN,
         REGION,
         ENABLE_VIRTUAL_HOST_STYLE,
+        DISABLE_EC2_METADATA,
     ]
     .contains(&key)
 }
@@ -82,6 +84,21 @@ pub fn build_s3_backend(
         }
     }
 
+    if let Some(disable_str) = connection.get(DISABLE_EC2_METADATA) {
+        let disable = disable_str.as_str().parse::<bool>().map_err(|e| {
+            error::InvalidConnectionSnafu {
+                msg: format!(
+                    "failed to parse the option {}={}, {}",
+                    DISABLE_EC2_METADATA, disable_str, e
+                ),
+            }
+            .build()
+        })?;
+        if disable {
+            builder = builder.disable_ec2_metadata();
+        }
+    }
+
     // TODO(weny): Consider finding a better way to eliminate duplicate code.
     Ok(ObjectStore::new(builder)
         .context(error::BuildBackendSnafu)?
@@ -109,6 +126,7 @@ mod tests {
         assert!(is_supported_in_s3(SESSION_TOKEN));
         assert!(is_supported_in_s3(REGION));
         assert!(is_supported_in_s3(ENABLE_VIRTUAL_HOST_STYLE));
+        assert!(is_supported_in_s3(DISABLE_EC2_METADATA));
         assert!(!is_supported_in_s3("foo"))
     }
 }

src/common/telemetry/src/metric.rs

@@ -71,6 +71,7 @@ pub fn convert_metric_to_write_request(
                         timestamp,
                     }],
                     exemplars: vec![],
+                    histograms: vec![],
                 }),
                 MetricType::GAUGE => timeseries.push(TimeSeries {
                     labels: convert_label(m.get_label(), mf_name, None),
@@ -79,6 +80,7 @@ pub fn convert_metric_to_write_request(
                         timestamp,
                     }],
                     exemplars: vec![],
+                    histograms: vec![],
                 }),
                 MetricType::HISTOGRAM => {
                     let h = m.get_histogram();
@@ -97,6 +99,7 @@ pub fn convert_metric_to_write_request(
                                 timestamp,
                             }],
                             exemplars: vec![],
+                            histograms: vec![],
                         });
                         if upper_bound.is_sign_positive() && upper_bound.is_infinite() {
                             inf_seen = true;
@@ -114,6 +117,7 @@ pub fn convert_metric_to_write_request(
                                 timestamp,
                             }],
                             exemplars: vec![],
+                            histograms: vec![],
                         });
                     }
                     timeseries.push(TimeSeries {
@@ -127,6 +131,7 @@ pub fn convert_metric_to_write_request(
                             timestamp,
                         }],
                         exemplars: vec![],
+                        histograms: vec![],
                     });
                     timeseries.push(TimeSeries {
                         labels: convert_label(
@@ -139,6 +144,7 @@ pub fn convert_metric_to_write_request(
                             timestamp,
                         }],
                         exemplars: vec![],
+                        histograms: vec![],
                     });
                 }
                 MetricType::SUMMARY => {
@@ -155,6 +161,7 @@ pub fn convert_metric_to_write_request(
                                 timestamp,
                             }],
                             exemplars: vec![],
+                            histograms: vec![],
                         });
                     }
                     timeseries.push(TimeSeries {
@@ -168,6 +175,7 @@ pub fn convert_metric_to_write_request(
                             timestamp,
                         }],
                         exemplars: vec![],
+                        histograms: vec![],
                     });
                     timeseries.push(TimeSeries {
                         labels: convert_label(
@@ -180,6 +188,7 @@ pub fn convert_metric_to_write_request(
                             timestamp,
                         }],
                         exemplars: vec![],
+                        histograms: vec![],
                     });
                 }
                 MetricType::UNTYPED => {
@@ -274,7 +283,7 @@ mod test {
 
         assert_eq!(
             format!("{:?}", write_quest.timeseries),
-            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }]"#
+            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }]"#
         );
 
         let gauge_opts = Opts::new("test_gauge", "test help")
@@ -288,7 +297,7 @@ mod test {
         let write_quest = convert_metric_to_write_request(mf, None, 0);
         assert_eq!(
             format!("{:?}", write_quest.timeseries),
-            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_gauge" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 42.0, timestamp: 0 }], exemplars: [] }]"#
+            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_gauge" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 42.0, timestamp: 0 }], exemplars: [], histograms: [] }]"#
         );
     }
 
@@ -305,20 +314,20 @@ mod test {
             .iter()
             .map(|x| format!("{:?}", x))
             .collect();
-        let ans = r#"TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.005" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.01" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.025" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.05" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.1" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.25" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "1" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "2.5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "10" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "+Inf" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_sum" }, Label { name: "a", value: "1" }], samples: [Sample { value: 0.25, timestamp: 0 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_count" }, Label { name: "a", value: "1" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }"#;
+        let ans = r#"TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.005" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.01" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.025" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.05" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.1" }], samples: [Sample { value: 0.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.25" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "0.5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "1" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "2.5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "5" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "10" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_bucket" }, Label { name: "a", value: "1" }, Label { name: "le", value: "+Inf" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_sum" }, Label { name: "a", value: "1" }], samples: [Sample { value: 0.25, timestamp: 0 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_count" }, Label { name: "a", value: "1" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }"#;
         assert_eq!(write_quest_str.join("\n"), ans);
     }
 
@@ -355,10 +364,10 @@ TimeSeries { labels: [Label { name: "__name__", value: "test_histogram_count" },
             .iter()
             .map(|x| format!("{:?}", x))
             .collect();
-        let ans = r#"TimeSeries { labels: [Label { name: "__name__", value: "test_summary" }, Label { name: "quantile", value: "50" }], samples: [Sample { value: 3.0, timestamp: 20 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_summary" }, Label { name: "quantile", value: "100" }], samples: [Sample { value: 5.0, timestamp: 20 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_summary_sum" }], samples: [Sample { value: 15.0, timestamp: 20 }], exemplars: [] }
-TimeSeries { labels: [Label { name: "__name__", value: "test_summary_count" }], samples: [Sample { value: 5.0, timestamp: 20 }], exemplars: [] }"#;
+        let ans = r#"TimeSeries { labels: [Label { name: "__name__", value: "test_summary" }, Label { name: "quantile", value: "50" }], samples: [Sample { value: 3.0, timestamp: 20 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_summary" }, Label { name: "quantile", value: "100" }], samples: [Sample { value: 5.0, timestamp: 20 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_summary_sum" }], samples: [Sample { value: 15.0, timestamp: 20 }], exemplars: [], histograms: [] }
+TimeSeries { labels: [Label { name: "__name__", value: "test_summary_count" }], samples: [Sample { value: 5.0, timestamp: 20 }], exemplars: [], histograms: [] }"#;
         assert_eq!(write_quest_str.join("\n"), ans);
     }
 
@@ -385,11 +394,11 @@ TimeSeries { labels: [Label { name: "__name__", value: "test_summary_count" }],
         let write_quest2 = convert_metric_to_write_request(mf, Some(&filter), 0);
         assert_eq!(
             format!("{:?}", write_quest1.timeseries),
-            r#"[TimeSeries { labels: [Label { name: "__name__", value: "filter_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [] }, TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 2.0, timestamp: 0 }], exemplars: [] }]"#
+            r#"[TimeSeries { labels: [Label { name: "__name__", value: "filter_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 1.0, timestamp: 0 }], exemplars: [], histograms: [] }, TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 2.0, timestamp: 0 }], exemplars: [], histograms: [] }]"#
         );
         assert_eq!(
             format!("{:?}", write_quest2.timeseries),
-            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 2.0, timestamp: 0 }], exemplars: [] }]"#
+            r#"[TimeSeries { labels: [Label { name: "__name__", value: "test_counter" }, Label { name: "a", value: "1" }, Label { name: "b", value: "2" }], samples: [Sample { value: 2.0, timestamp: 0 }], exemplars: [], histograms: [] }]"#
         );
     }
 }

src/mito2/src/gc.rs

@@ -540,7 +540,7 @@ impl LocalGcWorker {
     fn filter_deletable_files(
         &self,
         entries: Vec<Entry>,
-        in_use_filenames: &HashSet<&FileId>,
+        in_use_filenames: &HashSet<FileId>,
         may_linger_filenames: &HashSet<&FileId>,
         eligible_for_removal: &HashSet<&FileId>,
         unknown_file_may_linger_until: chrono::DateTime<chrono::Utc>,
@@ -641,9 +641,6 @@ impl LocalGcWorker {
             .flatten()
             .collect::<HashSet<_>>();
 
-        // in use filenames, include sst and index files
-        let in_use_filenames = in_used.iter().collect::<HashSet<_>>();
-
         // When full_file_listing is false, skip expensive list operations and only delete
         // files that are tracked in recently_removed_files
         if !self.full_file_listing {
@@ -653,7 +650,7 @@ impl LocalGcWorker {
             // 3. Have passed the lingering time
             let files_to_delete: Vec<FileId> = eligible_for_removal
                 .iter()
-                .filter(|file_id| !in_use_filenames.contains(*file_id))
+                .filter(|file_id| !in_used.contains(*file_id))
                 .map(|&f| *f)
                 .collect();
 
@@ -672,7 +669,7 @@ impl LocalGcWorker {
         let (all_unused_files_ready_for_delete, all_in_exist_linger_files) = self
             .filter_deletable_files(
                 all_entries,
-                &in_use_filenames,
+                in_used,
                 &may_linger_filenames,
                 &eligible_for_removal,
                 unknown_file_may_linger_until,

src/object-store/src/config.rs

@@ -16,7 +16,6 @@ use std::time::Duration;
 
 use common_base::readable_size::ReadableSize;
 use common_base::secrets::{ExposeSecret, SecretString};
-use common_telemetry::tracing::warn;
 use opendal::services::{Azblob, Gcs, Oss, S3};
 use serde::{Deserialize, Serialize};
 
@@ -118,23 +117,25 @@ pub struct S3Connection {
     /// By default, opendal will send API to https://s3.us-east-1.amazonaws.com/bucket_name
     /// Enabled, opendal will send API to https://bucket_name.s3.us-east-1.amazonaws.com
     pub enable_virtual_host_style: bool,
+    /// Disable EC2 metadata service.
+    /// By default, opendal will use EC2 metadata service to load credentials from the instance metadata,
+    /// when access key id and secret access key are not provided.
+    /// If enabled, opendal will *NOT* use EC2 metadata service.
+    pub disable_ec2_metadata: bool,
 }
 
 impl From<&S3Connection> for S3 {
     fn from(connection: &S3Connection) -> Self {
         let root = util::normalize_dir(&connection.root);
 
-        let mut builder = S3::default().root(&root).bucket(&connection.bucket);
+        let mut builder = S3::default()
+            .root(&root)
+            .bucket(&connection.bucket)
+            .access_key_id(connection.access_key_id.expose_secret())
+            .secret_access_key(connection.secret_access_key.expose_secret());
 
-        if !connection.access_key_id.expose_secret().is_empty()
-            && !connection.secret_access_key.expose_secret().is_empty()
-        {
-            builder = builder
-                .access_key_id(connection.access_key_id.expose_secret())
-                .secret_access_key(connection.secret_access_key.expose_secret());
-        } else {
-            warn!("No access key id or secret access key provided, using anonymous access");
-            builder = builder.allow_anonymous().disable_ec2_metadata();
+        if connection.disable_ec2_metadata {
+            builder = builder.disable_ec2_metadata();
         }
 
         if let Some(endpoint) = &connection.endpoint {

src/servers/src/prom_store.rs

@@ -26,7 +26,7 @@ use arrow::datatypes::{Float64Type, TimestampMillisecondType};
 use common_grpc::precision::Precision;
 use common_query::prelude::{greptime_timestamp, greptime_value};
 use common_recordbatch::{RecordBatch, RecordBatches};
-use common_telemetry::tracing;
+use common_telemetry::{tracing, warn};
 use datafusion::dataframe::DataFrame;
 use datafusion::prelude::{Expr, col, lit, regexp_match};
 use datafusion_common::ScalarValue;
@@ -415,6 +415,10 @@ pub fn to_grpc_row_insert_requests(request: &WriteRequest) -> Result<(RowInsertR
                 table_data.add_row(one_row);
             }
         }
+
+        if !series.histograms.is_empty() {
+            warn!("Native histograms are not supported yet, data ignored");
+        }
     }
 
     Ok(multi_table_data.into_row_insert_requests())

tests-integration/tests/http.rs

@@ -1586,6 +1586,7 @@ fn drop_lines_with_inconsistent_results(input: String) -> String {
         "endpoint =",
         "region =",
         "enable_virtual_host_style =",
+        "disable_ec2_metadata =",
         "cache_path =",
         "cache_capacity =",
         "memory_pool_size =",