Skip to content

[fix](recycler) Fix metrics statistics issues in recycler workflows #60697

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wyxxxcat wants to merge 1 commit into
base: master
Choose a base branch
from
+25 −15
Open

[fix](recycler) Fix metrics statistics issues in recycler workflows #60697

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
32 changes: 21 additions & 11 deletions cloud/src/recycler/recycler.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -3535,8 +3535,6 @@ int InstanceRecycler::delete_rowset_data(
metrics_context.total_recycled_num++;
}
});
segment_metrics_context_.report();
metrics_context.report();
}
return ret;
});
Expand All @@ -3554,8 +3552,6 @@ int InstanceRecycler::delete_rowset_data(
metrics_context.total_recycled_num++;
segment_metrics_context_.total_recycled_data_size += rs.total_disk_size();
segment_metrics_context_.total_recycled_num += rs.num_segments();
metrics_context.report();
segment_metrics_context_.report();
}
return ret;
});
Expand Down Expand Up @@ -3991,8 +3987,6 @@ int InstanceRecycler::recycle_tablet(int64_t tablet_id, RecyclerMetricsContext&
concurrent_delete_executor.add(
[&, rs_id = resource_id,
accessor_ptr = accessor_map_[resource_id]]() -> decltype(auto) {
std::unique_ptr<int, std::function<void(int*)>> defer(
(int*)0x01, [&](int*) { metrics_context.report(); });
int res = accessor_ptr->delete_directory(tablet_path_prefix(tablet_id));
if (res != 0) {
LOG(WARNING) << "failed to delete rowset data of tablet " << tablet_id
Expand Down Expand Up @@ -4596,8 +4590,6 @@ int InstanceRecycler::recycle_rowsets() {
segment_metrics_context_.total_recycled_data_size +=
rowset.rowset_meta().total_disk_size();
segment_metrics_context_.total_recycled_num += rowset.rowset_meta().num_segments();
segment_metrics_context_.report();
metrics_context.report();
return 0;
}

Expand Down Expand Up @@ -4721,6 +4713,11 @@ int InstanceRecycler::recycle_rowsets() {
num_recycled.fetch_add(async_recycled_rowset_keys.size(), std::memory_order_relaxed);
}
}

// Report final metrics after all concurrent tasks completed
segment_metrics_context_.report();
metrics_context.report();

return ret;
}

Expand Down Expand Up @@ -5125,6 +5122,11 @@ int InstanceRecycler::recycle_versioned_rowsets() {
num_recycled.fetch_add(async_recycled_rowset_keys.size(), std::memory_order_relaxed);
}
}

// Report final metrics after all concurrent tasks completed
segment_metrics_context_.report();
metrics_context.report();

return ret;
}

Expand Down Expand Up @@ -5468,6 +5470,11 @@ int InstanceRecycler::recycle_tmp_rowsets() {
std::move(loop_done));

worker_pool->stop();

// Report final metrics after all concurrent tasks completed
segment_metrics_context_.report();
metrics_context.report();

return ret;
}

Expand Down Expand Up @@ -5681,7 +5688,7 @@ int InstanceRecycler::recycle_expired_txn_label() {
const std::string task_name = "recycle_expired_txn_label";
int64_t num_scanned = 0;
int64_t num_expired = 0;
int64_t num_recycled = 0;
std::atomic_long num_recycled = 0;
RecyclerMetricsContext metrics_context(instance_id_, task_name);
int ret = 0;

Expand Down Expand Up @@ -5833,8 +5840,7 @@ int InstanceRecycler::recycle_expired_txn_label() {
LOG(WARNING) << "failed to delete expired txn, err=" << err << " key=" << hex(k);
return -1;
}
metrics_context.total_recycled_num = ++num_recycled;
metrics_context.report();
++num_recycled;

LOG(INFO) << "recycle expired txn, key=" << hex(k);
return 0;
Expand Down Expand Up @@ -5885,6 +5891,10 @@ int InstanceRecycler::recycle_expired_txn_label() {

ret = finished ? ret : -1;

// Update metrics after all concurrent tasks completed
metrics_context.total_recycled_num = num_recycled.load();
metrics_context.report();

TEST_SYNC_POINT_CALLBACK("InstanceRecycler::recycle_expired_txn_label.failure", &ret);

if (ret != 0) {
Expand Down
8 changes: 4 additions & 4 deletions cloud/src/recycler/recycler.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -192,6 +192,10 @@ class RecyclerMetricsContext {
g_bvar_recycler_instance_last_round_recycle_elpased_ts.put(
{instance_id, operation_type}, cost);
g_bvar_recycler_instance_recycle_round.put({instance_id, operation_type}, 1);
g_bvar_recycler_instance_recycle_total_bytes_since_started.put(
{instance_id, operation_type}, total_recycled_data_size.load());
g_bvar_recycler_instance_recycle_total_num_since_started.put(
{instance_id, operation_type}, total_recycled_num.load());
LOG(INFO) << "recycle instance: " << instance_id
<< ", operation type: " << operation_type << ", cost: " << cost
<< " ms, total recycled num: " << total_recycled_num.load()
Expand Down Expand Up @@ -222,12 +226,8 @@ class RecyclerMetricsContext {
} else {
g_bvar_recycler_instance_last_round_recycled_bytes.put(
{instance_id, operation_type}, total_recycled_data_size.load());
g_bvar_recycler_instance_recycle_total_bytes_since_started.put(
{instance_id, operation_type}, total_recycled_data_size.load());
g_bvar_recycler_instance_last_round_recycled_num.put({instance_id, operation_type},
total_recycled_num.load());
g_bvar_recycler_instance_recycle_total_num_since_started.put(
{instance_id, operation_type}, total_recycled_num.load());
}
}
}
Expand Down