Files
impala/common/thrift/ExecStats.thrift
Nong Li 5d903efca3 ExecSummary
The runtime profile as we present it is not very useful and I think the structure of
it makes it hard to consume. This patch adds a new client facing schemed set of
counters that are collected from the runtime profiles. For example, with this structure
it would be easy to have the shell get the stats of a running query and print a useful
progress report or to check the most relevant metrics for diagnosing issues.

Here's an example of the output for one of the tpch queries:
Operator              #Hosts   Avg Time   Max Time    #Rows  Est. #Rows  Peak Mem  Est. Peak Mem  Detail
------------------------------------------------------------------------------------------------------------------------
09:MERGING-EXCHANGE        1   79.738us   79.738us        5           5         0        -1.00 B  UNPARTITIONED
05:TOP-N                   3   84.693us   88.810us        5           5  12.00 KB       120.00 B
04:AGGREGATE               3    5.263ms    6.432ms        5           5  44.00 KB       10.00 MB  MERGE FINALIZE
08:AGGREGATE               3   16.659ms   27.444ms   52.52K     600.12K   3.20 MB       15.11 MB  MERGE
07:EXCHANGE                3    2.644ms      5.1ms   52.52K     600.12K         0              0  HASH(o_orderpriority)
03:AGGREGATE               3  342.913ms  966.291ms   52.52K     600.12K  10.80 MB       15.11 MB
02:HASH JOIN               3    2s165ms    2s171ms  144.87K     600.12K  13.63 MB      941.01 KB  INNER JOIN, BROADCAST
|--06:EXCHANGE             3    8.296ms    8.692ms   57.22K      15.00K         0              0  BROADCAST
|  01:SCAN HDFS            2    1s412ms    1s978ms   57.22K      15.00K  24.21 MB      176.00 MB  tpch.orders o
00:SCAN HDFS               3    8s032ms    8s558ms    3.79M     600.12K  32.29 MB      264.00 MB  tpch.lineitem l

Change-Id: Iaad4b9dd577c375006313f19442bee6d3e27246a
Reviewed-on: http://gerrit.ent.cloudera.com:8080/2964
Reviewed-by: Nong Li <nong@cloudera.com>
Tested-by: jenkins
2014-06-11 03:10:11 -07:00

96 lines
3.0 KiB
Thrift

// Copyright 2012 Cloudera Inc.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.
namespace cpp impala
namespace java com.cloudera.impala.thrift
include "Status.thrift"
include "Types.thrift"
enum TExecState {
REGISTERED = 0,
PLANNING = 1,
QUEUED = 2,
RUNNING = 3,
FINISHED = 4,
CANCELLED = 5,
FAILED = 6,
}
// Execution stats for a single plan node.
struct TExecStats {
// The wall clock time spent on the "main" thread. This is the user perceived
// latency. This value indicates the current bottleneck.
// Note: anywhere we have a queue between operators, this time can fluctuate
// significantly without the overall query time changing much (i.e. the bottleneck
// moved to another operator). This is unavoidable though.
1: optional i64 latency_ns
// Total CPU time spent across all threads. For operators that have an async
// component (e.g. multi-threaded) this will be >= latency_ns.
2: optional i64 cpu_time_ns
// Number of rows returned.
3: optional i64 cardinality
// Peak memory used (in bytes).
4: optional i64 memory_used
}
// Summary for a single plan node. This includes labels for how to display the
// node as well as per instance stats.
struct TPlanNodeExecSummary {
1: required Types.TPlanNodeId node_id
2: required i32 fragment_id
3: required string label
4: optional string label_detail
5: required i32 num_children
// Estimated stats generated by the planner
6: optional TExecStats estimated_stats
// One entry for each BE executing this plan node.
7: optional list<TExecStats> exec_stats
// One entry for each BE executing this plan node. True if this plan node is still
// running.
8: optional list<bool> is_active
// If true, this plan node is an exchange node that is the receiver of a broadcast.
9: optional bool is_broadcast
}
// Execution summary of an entire query.
struct TExecSummary {
// State of the query.
1: required TExecState state
// Contains the error if state is FAILED.
2: optional Status.TStatus status
// Flattened execution summary of the plan tree.
3: optional list<TPlanNodeExecSummary> nodes
// For each exch node in 'nodes', contains the index to the root node of the sending
// fragment for this exch. Both the key and value are indices into 'nodes'.
4: optional map<i32, i32> exch_to_sender_map
// List of errors that were encountered during execution. This can be non-empty
// even if status is okay, in which case it contains errors that impala skipped
// over.
5: optional list<string> error_logs
}