Skip to content

Optimize Bucket configs for Histogram. #798

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

Closed
xiaoxichen wants to merge 1 commit into from
Closed

Optimize Bucket configs for Histogram. #798

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
4 changes: 2 additions & 2 deletions conanfile.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

class HomestoreConan(ConanFile):
name = "homestore"
version = "6.20.19"
version = "6.20.20"

homepage = "https://github.com/eBay/Homestore"
description = "HomeStore Storage Engine"
Expand Down Expand Up @@ -53,7 +53,7 @@ def build_requirements(self):

def requirements(self):
self.requires("iomgr/[^11.3]@oss/master", transitive_headers=True)
self.requires("sisl/[^12.2]@oss/master", transitive_headers=True)
self.requires("sisl/[^12.4.7]@oss/master", transitive_headers=True)
self.requires("nuraft_mesg/[~3.8.7]@oss/main", transitive_headers=True)

self.requires("farmhash/cci.20190513@", transitive_headers=True)
Expand Down
15 changes: 9 additions & 6 deletions src/include/homestore/btree/detail/btree_internal.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -315,9 +315,9 @@ class BtreeMetrics : public sisl::MetricsGroup {
REGISTER_COUNTER(btree_num_pc_gen_mismatch, "Number of gen mismatches to recover");

REGISTER_HISTOGRAM(btree_int_node_occupancy, "Interior node occupancy", "btree_node_occupancy",
{"node_type", "interior"}, HistogramBucketsType(LinearUpto128Buckets));
{"node_type", "interior"}, HistogramBucketsType(PercentileBuckets));
REGISTER_HISTOGRAM(btree_leaf_node_occupancy, "Leaf node occupancy", "btree_node_occupancy",
{"node_type", "leaf"}, HistogramBucketsType(LinearUpto128Buckets));
{"node_type", "leaf"}, HistogramBucketsType(PercentileBuckets));
REGISTER_COUNTER(btree_retry_count, "number of retries");
REGISTER_COUNTER(write_err_cnt, "number of errors in write");
REGISTER_COUNTER(query_err_cnt, "number of errors in query");
Expand All @@ -326,13 +326,16 @@ class BtreeMetrics : public sisl::MetricsGroup {
REGISTER_COUNTER(btree_remove_ops_count, "number of btree operations");
REGISTER_HISTOGRAM(btree_exclusive_time_in_int_node,
"Exclusive time spent (Write locked) on interior node (ns)", "btree_exclusive_time_in_node",
{"node_type", "interior"});
{"node_type", "interior"}, HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(btree_exclusive_time_in_leaf_node, "Exclusive time spent (Write locked) on leaf node (ns)",
"btree_exclusive_time_in_node", {"node_type", "leaf"});
"btree_exclusive_time_in_node", {"node_type", "leaf"},
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(btree_inclusive_time_in_int_node, "Inclusive time spent (Read locked) on interior node (ns)",
"btree_inclusive_time_in_node", {"node_type", "interior"});
"btree_inclusive_time_in_node", {"node_type", "interior"},
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(btree_inclusive_time_in_leaf_node, "Inclusive time spent (Read locked) on leaf node (ns)",
"btree_inclusive_time_in_node", {"node_type", "leaf"});
"btree_inclusive_time_in_node", {"node_type", "leaf"},
HistogramBucketsType(OpLatecyBuckets));

register_me_to_farm();
}
Expand Down
2 changes: 1 addition & 1 deletion src/include/homestore/checkpoint/cp_mgr.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@ class CPMgrMetrics : public sisl::MetricsGroup {
explicit CPMgrMetrics() : sisl::MetricsGroup("CPMgr") {
REGISTER_COUNTER(back_to_back_cps, "back to back cp");
REGISTER_COUNTER(cp_cnt, "cp cnt");
REGISTER_HISTOGRAM(cp_latency, "cp latency (in us)");
REGISTER_HISTOGRAM(cp_latency, "cp latency (in us)", HistogramBucketsType(OpLatecyBuckets));
register_me_to_farm();
}

Expand Down
3 changes: 2 additions & 1 deletion src/include/homestore/meta_service.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -62,7 +62,8 @@ class MetablkMetrics : public sisl::MetricsGroupWrapper {
REGISTER_COUNTER(compress_backoff_memory_cnt, "compression back-off cnt because of exceending memory limit")
REGISTER_COUNTER(compress_backoff_ratio_cnt, "compression back-off cnt because of exceeding ratio limit");

REGISTER_HISTOGRAM(compress_ratio_percent, "compression ration percentage");
REGISTER_HISTOGRAM(compress_ratio_percent, "compression ration percentage",
HistogramBucketsType(PercentileBuckets));
register_me_to_farm();
}

Expand Down
2 changes: 1 addition & 1 deletion src/lib/blkalloc/varsize_blk_allocator.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -180,7 +180,7 @@ class BlkAllocMetrics : public sisl::MetricsGroup {
REGISTER_COUNTER(num_blks_alloc_direct, "Number of blks alloc attempt directly because of empty cache");

REGISTER_HISTOGRAM(frag_pct_distribution, "Distribution of fragmentation percentage",
HistogramBucketsType(LinearUpto64Buckets));
HistogramBucketsType(PercentileBuckets));

register_me_to_farm();
}
Expand Down
6 changes: 4 additions & 2 deletions src/lib/device/physical_dev.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -48,8 +48,10 @@ class PhysicalDevMetrics : public sisl::MetricsGroupWrapper {
REGISTER_COUNTER(drive_spurios_events, "Total number of spurious events per drive");
REGISTER_COUNTER(drive_skipped_chunk_bm_writes, "Total number of skipped writes for chunk bitmap");

REGISTER_HISTOGRAM(drive_write_latency, "BlkStore drive write latency in us");
REGISTER_HISTOGRAM(drive_read_latency, "BlkStore drive read latency in us");
REGISTER_HISTOGRAM(drive_write_latency, "BlkStore drive write latency in us",
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(drive_read_latency, "BlkStore drive read latency in us",
HistogramBucketsType(OpLatecyBuckets));

REGISTER_HISTOGRAM(write_io_sizes, "Write IO Sizes", "io_sizes", {"io_direction", "write"},
HistogramBucketsType(ExponentialOfTwoBuckets));
Expand Down
3 changes: 2 additions & 1 deletion src/lib/device/virtual_dev.hpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -54,7 +54,8 @@ class VirtualDevMetrics : public sisl::MetricsGroupWrapper {
REGISTER_COUNTER(default_chunk_allocation_cnt, "default chunk allocation count");
REGISTER_COUNTER(random_chunk_allocation_cnt,
"random chunk allocation count"); // ideally it should be zero for hdd
REGISTER_HISTOGRAM(blk_alloc_latency, "Blk allocation latency", "blk_alloc_latency");
REGISTER_HISTOGRAM(blk_alloc_latency, "Blk allocation latency", "blk_alloc_latency",
HistogramBucketsType(OpLatecyBuckets));
register_me_to_farm();
}

Expand Down
14 changes: 8 additions & 6 deletions src/lib/logstore/log_store_service.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -404,23 +404,25 @@ LogStoreServiceMetrics::LogStoreServiceMetrics() : sisl::MetricsGroup("LogStores
{"op", "write"});
REGISTER_COUNTER(logstore_read_count, "Total number of read requests to log stores", "logstore_op_count",
{"op", "read"});
REGISTER_HISTOGRAM(logstore_append_latency, "Logstore append latency", "logstore_op_latency", {"op", "write"});
REGISTER_HISTOGRAM(logstore_append_latency, "Logstore append latency", "logstore_op_latency", {"op", "write"},
HistogramBucketsType(OpLatecyBuckets));
#ifdef _PRERELEASE
REGISTER_HISTOGRAM(logstore_stream_tracker_lock_latency, "Logstore stream tracker lock latency",
"logstore_stream_tracker_lock_latency");
#endif
REGISTER_HISTOGRAM(logstore_read_latency, "Logstore read latency", "logstore_op_latency", {"op", "read"});
REGISTER_HISTOGRAM(logstore_read_latency, "Logstore read latency", "logstore_op_latency", {"op", "read"},
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(logdev_flush_size_distribution, "Distribution of flush data size",
HistogramBucketsType(ExponentialOfTwoBuckets));
REGISTER_HISTOGRAM(logdev_flush_records_distribution, "Distribution of num records to flush",
HistogramBucketsType(LinearUpto128Buckets));
REGISTER_HISTOGRAM(logstore_record_size, "Distribution of log record size",
HistogramBucketsType(ExponentialOfTwoBuckets));
REGISTER_HISTOGRAM(logdev_flush_done_msg_time_ns, "Logdev flush completion msg time in ns");
REGISTER_HISTOGRAM(logdev_post_flush_processing_latency,
"Logdev post flush processing (including callbacks) latency");
REGISTER_HISTOGRAM(logdev_flush_time_us, "time elapsed since last flush time in us");
REGISTER_HISTOGRAM(logdev_fsync_time_us, "Logdev fsync completion time in us");
"Logdev post flush processing (including callbacks) latency",
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(logdev_flush_time_us, "time elapsed since last flush time in us",
HistogramBucketsType(OpLatecyBuckets));

register_me_to_farm();
}
Expand Down
29 changes: 11 additions & 18 deletions src/lib/replication/repl_dev/raft_repl_dev.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -90,42 +90,35 @@ class RaftReplDevMetrics : public sisl::MetricsGroup {
// leader: data write latency;
// follower: from rreq push data received to data write completion;
REGISTER_HISTOGRAM(rreq_data_write_latency_us, "rreq data write latency in us", "rreq_data_op_latency",
{"op", "write"});
REGISTER_HISTOGRAM(rreq_data_read_latency_us, "rreq data read latency in us", "rreq_data_op_latency",
{"op", "read"}); // placeholder
{"op", "write"}, HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(rreq_push_data_latency_us, "rreq data write latency in us", "rreq_data_op_latency",
{"op", "push"});
// latency from req received to sending response
REGISTER_HISTOGRAM(rreq_data_write_respond_latency_us, "rreq data write and respond latency in us",
"rreq_data_op_latency", {"op", "respond"});
// latency from req received to rpc complete
REGISTER_HISTOGRAM(rreq_data_write_complete_latency_us, "rreq data rpc complete latency in us",
"rreq_data_op_latency", {"op", "complete"});
{"op", "push"}, HistogramBucketsType(OpLatecyBuckets));
// latency from follower->originator->follower, not including actual data write on follower;
REGISTER_HISTOGRAM(rreq_data_fetch_latency_us, "rreq data fetch latency in us", "rreq_data_op_latency",
{"op", "fetch"});
{"op", "fetch"}, HistogramBucketsType(OpLatecyBuckets));

/* from rreq creation to data ops completion */
REGISTER_HISTOGRAM(rreq_total_data_read_latency_us, "rreq data read latency in us", "rdev_data_op_latency",
{"op", "read"}); // placeholder
REGISTER_HISTOGRAM(rreq_total_data_write_latency_us, "rreq data write latency in us", "rdev_data_op_latency",
{"op", "write"});
{"op", "write"}, HistogramBucketsType(OpLatecyBuckets));

REGISTER_HISTOGRAM(rreq_pieces_per_write, "Number of individual pieces per write",
HistogramBucketsType(LinearUpto64Buckets));
HistogramBucketsType(SteppedUpto32Buckets));

// In the identical layout chunk, the blk num of the follower and leader is expected to be the same.
// However, due to the concurrency between the data channel and the raft channel, there might be some
// allocation differences on the same lsn. When a leader switch occurs, these differences could become garbage.
// This metric can partially reflect the potential amount of garbage.
REGISTER_HISTOGRAM(blk_diff_with_proposer,
"allocated blk num diff on the same lsn with proposer when chunk usage >= 0.9");
"allocated blk num diff on the same lsn with proposer when chunk usage >= 0.9",
HistogramBucketsType(ExponentialOfTwoBuckets));

// Raft channel metrics
REGISTER_HISTOGRAM(raft_end_of_append_batch_latency_us, "Raft end_of_append_batch latency in us",
"raft_logstore_append_latency", {"op", "end_of_append_batch"});
"raft_logstore_append_latency", {"op", "end_of_append_batch"},
HistogramBucketsType(OpLatecyBuckets));
REGISTER_HISTOGRAM(data_channel_wait_latency_us, "Data channel wait latency in us",
"raft_logstore_append_latency", {"op", "wait_for_data"});
"raft_logstore_append_latency", {"op", "wait_for_data"},
HistogramBucketsType(OpLatecyBuckets));

register_me_to_farm();
}
Expand Down
Loading