| <?xml version="1.0" encoding="UTF-8"?> |
| <!-- |
| Licensed to the Apache Software Foundation (ASF) under one |
| or more contributor license agreements. See the NOTICE file |
| distributed with this work for additional information |
| regarding copyright ownership. The ASF licenses this file |
| to you 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. |
| --> |
| <!DOCTYPE concept PUBLIC "-//OASIS//DTD DITA Concept//EN" "concept.dtd"> |
| <concept id="workload_mgmt"> |
| |
| <title>Impala Workload Management</title> |
| |
| <titlealts audience="PDF"> |
| |
| <navtitle>Workload Management</navtitle> |
| |
| </titlealts> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="Impala"/> |
| <data name="Category" value="Workload Management"/> |
| <data name="Category" value="Administrators"/> |
| <data name="Category" value="Data Analysts"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| |
| <p> |
| Impala Workload Management provides visibility into running and completed queries, |
| which can be accessed through any SQL client. |
| </p> |
| |
| <p outputclass="toc inpage"/> |
| </conbody> |
| |
| <concept id="wm_how_it_works"> |
| |
| <title>How It Works</title> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="How It Works"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| <p> |
| As queries progress through their lifecycle, telemetry is collected and tracked by |
| the coordinator responsible for executing each query. Workload Management |
| maintains this telemetry using a dedicated in-memory data structure. Retrieving the |
| telemetry for all queries currently running and recently completed is achieved |
| by querying the <codeph>sys.impala_query_live</codeph> table. This table is not a |
| traditional table since it is not backed by files on disk but by in-memory data |
| structures on each coordinator. |
| </p> |
| |
| <p> |
| At regular intervals, based on the <codeph>query_log_write_interval_s</codeph> |
| or <codeph>query_log_max_queued</codeph> startup flags, the Workload Management |
| processing loop wakes up to move telemetry for completed queries to the Apache |
| Iceberg table <codeph>sys.impala_query_log</codeph>. This table is backed by |
| persistent storage. The Workload Management processing loop runs one final time as |
| the last step of a graceful shutdown. |
| </p> |
| |
| <p> |
| Because the telemetry move is accomplished by an <codeph>INSERT</codeph> DML statement, |
| at least one active Impala executor must be available in the cluster. |
| </p> |
| |
| <p> |
| If an error occurs while the telemetry move <codeph>INSERT</codeph> DML is executing, |
| the telemetry is re-queued in the coordinator's in-memory structure to be retried |
| during the next processing cycle. Each query's telemetry is attempted up to three |
| times before it is dropped. |
| </p> |
| |
| </conbody> |
| </concept> |
| |
| <concept id="wm_table_structure"> |
| |
| <title>Table Structure</title> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="Table Schema"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| |
| <p> |
| Both <codeph>sys.impala_query_live</codeph> and <codeph>sys.impala_query_log</codeph> |
| share the same schema. The columns in these tables capture detailed metadata about |
| each query, including session information, query options, performance metrics, |
| admission status, memory usage, and more. |
| </p> |
| |
| <p> |
| The following table describes the columns available in these tables: |
| </p> |
| |
| <table> |
| <tgroup cols="4" rowsep="1" colsep="1"> |
| <colspec colname="1" colwidth = "6*" /> |
| <colspec colname="1" colwidth = "3*" /> |
| <colspec colname="1" colwidth = "20*" /> |
| <colspec colname="1" colwidth = "10*" /> |
| <thead> |
| <row> |
| <entry>Column Name</entry> |
| <entry>Type</entry> |
| <entry>Description</entry> |
| <entry>Sample Value</entry> |
| </row> |
| </thead> |
| <tbody> |
| <row> |
| <entry>cluster_id</entry> |
| <entry>String</entry> |
| <entry> |
| Value of the <codeph>cluster_id</codeph> coordinator startup flag. |
| </entry> |
| <entry><codeph>cluster-123</codeph></entry> |
| </row> |
| <row> |
| <entry>query_id</entry> |
| <entry>String</entry> |
| <entry> |
| Impala assigned query identifier. |
| </entry> |
| <entry><codeph>214d08bef0831e7a:3c65392400000000</codeph></entry> |
| </row> |
| <row> |
| <entry>session_id</entry> |
| <entry>String</entry> |
| <entry> |
| Impala assigned session identifier. |
| </entry> |
| <entry><codeph>ea4f661af43993d8:587839553a41adb8</codeph></entry> |
| </row> |
| <row> |
| <entry>session_type</entry> |
| <entry>String</entry> |
| <entry> |
| Type of the client session. |
| </entry> |
| <entry><codeph>HIVESERVER2</codeph></entry> |
| </row> |
| <row> |
| <entry>hiveserver2_protocol_version</entry> |
| <entry>String</entry> |
| <entry> |
| HiveServer2 protocol version used by the client. |
| </entry> |
| <entry><codeph>V6</codeph></entry> |
| </row> |
| <row> |
| <entry>db_user</entry> |
| <entry>String</entry> |
| <entry> |
| Effective cluster user. |
| </entry> |
| <entry><codeph>csso_name</codeph></entry> |
| </row> |
| <row> |
| <entry>db_user_connection</entry> |
| <entry>String</entry> |
| <entry> |
| Username from the authenticated client. |
| </entry> |
| <entry><codeph>csso_name</codeph></entry> |
| </row> |
| <row> |
| <entry>db_name</entry> |
| <entry>String</entry> |
| <entry> |
| Database queried. |
| </entry> |
| <entry><codeph>default</codeph></entry> |
| </row> |
| <row> |
| <entry>impala_coordinator</entry> |
| <entry>String</entry> |
| <entry> |
| Host and port of the coordinator for the query. |
| </entry> |
| <entry><codeph>coord-22899:27000</codeph></entry> |
| </row> |
| <row> |
| <entry>query_status</entry> |
| <entry>String</entry> |
| <entry> |
| Final status of the query. |
| </entry> |
| <entry><codeph>OK</codeph></entry> |
| </row> |
| <row> |
| <entry>query_state</entry> |
| <entry>String</entry> |
| <entry> |
| Final state of the query. |
| </entry> |
| <entry><codeph>FINISHED</codeph></entry> |
| </row> |
| <row> |
| <entry>impala_query_end_state</entry> |
| <entry>String</entry> |
| <entry> |
| Final Impala specific state of the query. |
| </entry> |
| <entry><codeph>FINISHED</codeph></entry> |
| </row> |
| <row> |
| <entry>query_type</entry> |
| <entry>String</entry> |
| <entry> |
| Type of the query (e.g. <codeph>QUERY</codeph>, <codeph>DDL</codeph>). |
| </entry> |
| <entry><codeph>QUERY</codeph></entry> |
| </row> |
| <row> |
| <entry>network_address</entry> |
| <entry>String</entry> |
| <entry> |
| Client IP and port. |
| </entry> |
| <entry><codeph>127.0.0.1:40120</codeph></entry> |
| </row> |
| <row> |
| <entry>start_time_utc</entry> |
| <entry>Timestamp</entry> |
| <entry> |
| Time the query started (UTC). |
| </entry> |
| <entry><codeph>2025-03-28 17:13:46.414316000</codeph></entry> |
| </row> |
| <row> |
| <entry>total_time_ms</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Total time between start and end, in milliseconds. |
| </entry> |
| <entry><codeph>136.121</codeph></entry> |
| </row> |
| <row> |
| <entry>query_opts_config</entry> |
| <entry>String</entry> |
| <entry> |
| Query options as comma-separated key-value pairs. |
| </entry> |
| <entry><codeph>TIMEZONE=America/Los_Angeles,QUERY_TIMEOUT_S=600,MT_DOP=12</codeph></entry> |
| </row> |
| <row> |
| <entry>resource_pool</entry> |
| <entry>String</entry> |
| <entry> |
| Name of the resource pool used. |
| </entry> |
| <entry><codeph>default-pool</codeph></entry> |
| </row> |
| <row> |
| <entry>per_host_mem_estimate</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Per-host memory estimate, in bytes. |
| </entry> |
| <entry><codeph>5</codeph></entry> |
| </row> |
| <row> |
| <entry>dedicated_coord_mem_estimate</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Coordinator memory estimate, in bytes. |
| </entry> |
| <entry><codeph>4</codeph></entry> |
| </row> |
| <row> |
| <entry>per_host_fragment_instances</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated host and fragment instance count. |
| </entry> |
| <entry><codeph>myhost-1:27000=1,myhost-2:27001=2</codeph></entry> |
| </row> |
| <row> |
| <entry>backends_count</entry> |
| <entry>Integer</entry> |
| <entry> |
| Count of the number of backends used by the query. |
| </entry> |
| <entry><codeph>2</codeph></entry> |
| </row> |
| <row> |
| <entry>admission_result</entry> |
| <entry>String</entry> |
| <entry> |
| Admission control result. Not applicable to <codeph>DDLs</codeph>. |
| </entry> |
| <entry><codeph>Admitted immediately</codeph></entry> |
| </row> |
| <row> |
| <entry>cluster_memory_admitted</entry> |
| <entry>Integer</entry> |
| <entry> |
| Admitted cluster memory in bytes. |
| </entry> |
| <entry><codeph>4</codeph></entry> |
| </row> |
| <row> |
| <entry>executor_group</entry> |
| <entry>String</entry> |
| <entry> |
| Executor group name. |
| </entry> |
| <entry><codeph>executor_group</codeph></entry> |
| </row> |
| <row> |
| <entry>executor_groups</entry> |
| <entry>String</entry> |
| <entry> |
| All considered executor groups (including rejected ones). |
| </entry> |
| <entry><codeph>executor_group1,executor_group2</codeph></entry> |
| </row> |
| <row> |
| <entry>exec_summary</entry> |
| <entry>String</entry> |
| <entry> |
| Full executor summary text. |
| </entry> |
| <entry></entry> |
| </row> |
| <row> |
| <entry>num_rows_fetched</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Number of rows fetched. |
| </entry> |
| <entry><codeph>6001215</codeph></entry> |
| </row> |
| <row> |
| <entry>row_materialization_rows_per_sec</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Rows materialized per second. |
| </entry> |
| <entry><codeph>3780</codeph></entry> |
| </row> |
| <row> |
| <entry>row_materialization_time_ms</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Time spent materializing rows, in milliseconds. |
| </entry> |
| <entry><codeph>1.58</codeph></entry> |
| </row> |
| <row> |
| <entry>compressed_bytes_spilled</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Bytes spilled to scratch disk. |
| </entry> |
| <entry><codeph>241515</codeph></entry> |
| </row> |
| <row> |
| <entry>event_planning_finished</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>27.253</codeph></entry> |
| </row> |
| <row> |
| <entry>event_submit_for_admission</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>30.204</codeph></entry> |
| </row> |
| <row> |
| <entry>event_completed_admission</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>30.986</codeph></entry> |
| </row> |
| <row> |
| <entry>event_all_backends_started</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>31.969</codeph></entry> |
| </row> |
| <row> |
| <entry>event_rows_available</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>31.969</codeph></entry> |
| </row> |
| <row> |
| <entry>event_first_row_fetched</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>135.175</codeph></entry> |
| </row> |
| <row> |
| <entry>event_last_row_fetched</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>135.181</codeph></entry> |
| </row> |
| <row> |
| <entry>event_unregister_query</entry> |
| <entry>Decimal(18,3)</entry> |
| <entry> |
| Timeline event (milliseconds since the query was received). |
| </entry> |
| <entry><codeph>141.435</codeph></entry> |
| </row> |
| <row> |
| <entry>read_io_wait_total_ms</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Total read I/O wait time, in milliseconds. |
| </entry> |
| <entry><codeph>15.091</codeph></entry> |
| </row> |
| <row> |
| <entry>read_io_wait_mean_ms</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Average read I/O wait time of all executors, in milliseconds. |
| </entry> |
| <entry><codeph>35.515</codeph></entry> |
| </row> |
| <row> |
| <entry>bytes_read_cache_total</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Total bytes read from data cache. |
| </entry> |
| <entry><codeph>45823</codeph></entry> |
| </row> |
| <row> |
| <entry>bytes_read_total</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Total bytes read. |
| </entry> |
| <entry><codeph>745227</codeph></entry> |
| </row> |
| <row> |
| <entry>pernode_peak_mem_min</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Min peak memory usage across nodes. |
| </entry> |
| <entry><codeph>5552846</codeph></entry> |
| </row> |
| <row> |
| <entry>pernode_peak_mem_max</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Max peak memory usage across nodes. |
| </entry> |
| <entry><codeph>5552846</codeph></entry> |
| </row> |
| <row> |
| <entry>pernode_peak_mem_mean</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Mean peak memory usage across nodes. |
| </entry> |
| <entry><codeph>5552846</codeph></entry> |
| </row> |
| <row> |
| <entry>sql</entry> |
| <entry>String</entry> |
| <entry> |
| User-provided SQL, redacted by Impala. |
| </entry> |
| <entry><codeph>SELECT a.col1, a.col2 from db.tbl a</codeph></entry> |
| </row> |
| <row> |
| <entry>plan</entry> |
| <entry>String</entry> |
| <entry> |
| Full query plan text. |
| </entry> |
| <entry></entry> |
| </row> |
| <row> |
| <entry>tables_queried</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated string containing all the tables queried in the SQL |
| statement. Aliased tables are resolved to their actual table names. |
| </entry> |
| <entry><codeph>db.tbl</codeph></entry> |
| </row> |
| <row> |
| <entry>select_columns</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated fully qualified table names. |
| </entry> |
| <entry><codeph>db.tbl.col1,db.tbl.col2</codeph></entry> |
| </row> |
| <row> |
| <entry>where_columns</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated list of fully qualified column names from the |
| SQL where clause. Aliased columns are resolved to their actual column names. |
| </entry> |
| <entry><codeph>db.tbl.col1,db.tbl.col2</codeph></entry> |
| </row> |
| <row> |
| <entry>join_columns</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated list of fully qualified column names from the |
| SQL <codeph>JOIN</codeph> clauses. Aliased columns are resolved to their |
| actual column names. |
| </entry> |
| <entry><codeph>db.tbl.col1,db.tbl.col2</codeph></entry> |
| </row> |
| <row> |
| <entry>aggregate_columns</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated list of fully qualified column names from the |
| SQL <codeph>GROUP BY</codeph> and <codeph>HAVING</codeph> clauses. Aliased |
| columns are resolved to their actual column names. |
| </entry> |
| <entry><codeph>db.tbl.col1,db.tbl.col2</codeph></entry> |
| </row> |
| <row> |
| <entry>orderby_columns</entry> |
| <entry>String</entry> |
| <entry> |
| Comma-separated list of fully qualified column names from the |
| SQL <codeph>ORDER BY</codeph> clause. Aliased columns are resolved to their |
| actual column names. |
| </entry> |
| <entry><codeph>db.tbl.col1,db.tbl.col2</codeph></entry> |
| </row> |
| <row> |
| <entry>coordinator_slots</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Number of coordinator query slots used. |
| </entry> |
| <entry><codeph>1</codeph></entry> |
| </row> |
| <row> |
| <entry>executor_slots</entry> |
| <entry>Bigint</entry> |
| <entry> |
| Number of executor query slots used by a single executor. |
| </entry> |
| <entry><codeph>1</codeph></entry> |
| </row> |
| </tbody> |
| </tgroup> |
| </table> |
| |
| </conbody> |
| </concept> |
| |
| <concept id="wm_setup"> |
| |
| <title>Setup</title> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="Data Setup"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| |
| <p> |
| Workload Management is configured by startup flags. The only required flag is |
| <codeph>--enable_workload_management=true</codeph>, which must be set on all |
| coordinator and catalogd processes. Other coordinator flags provide flexibility to |
| tailor Workload Management to specific needs. The table below lists the common |
| flags that may need to be modified. |
| </p> |
| |
| <table> |
| <tgroup cols="4" rowsep="1" colsep="1"> |
| <colspec colname="1" colwidth = "10*" /> |
| <colspec colname="1" colwidth = "5*" /> |
| <colspec colname="1" colwidth = "5*" /> |
| <colspec colname="1" colwidth = "34*" /> |
| <thead> |
| <row> |
| <entry>Flag Name</entry> |
| <entry>Type</entry> |
| <entry>Default</entry> |
| <entry>Description</entry> |
| </row> |
| </thead> |
| <tbody> |
| <row> |
| <entry>enable_workload_management</entry> |
| <entry>Boolean</entry> |
| <entry>false</entry> |
| <entry>Enables Workload Management when set to <codeph>true</codeph></entry> |
| </row> |
| <row> |
| <entry>cluster_id</entry> |
| <entry>String</entry> |
| <entry>(empty)</entry> |
| <entry> |
| Identifies the current cluster. This value is written to |
| the <codeph>cluster_id</codeph> column in the Workload Management table. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_write_interval_s</entry> |
| <entry>Integer</entry> |
| <entry>300</entry> |
| <entry> |
| Number of seconds to wait before batching and inserting completed queries |
| into the query history table. Helps reduce small file creation by batching |
| multiple records into a single insert. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_max_queued</entry> |
| <entry>Integer</entry> |
| <entry>3000</entry> |
| <entry> |
| Maximum number of completed queries that can be queued before |
| being written to the query history table. This limit is independent of |
| <codeph>query_log_write_interval_s</codeph>. If the queue reaches this size, |
| the batch is flushed immediately. A value of <codeph>0</codeph> disables this limit. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_max_sql_length</entry> |
| <entry>Integer</entry> |
| <entry>16777216</entry> |
| <entry> |
| Maximum length of a SQL statement to be recorded in the |
| <codeph>sys.impala_query_log</codeph> table. Longer statements are truncated. |
| Escaped characters count toward this limit. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_max_plan_length</entry> |
| <entry>Integer</entry> |
| <entry>16777216</entry> |
| <entry> |
| Maximum length of a query plan to be recorded in the query |
| log. Plans longer than this value are truncated. Escaped characters count |
| toward the limit. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_request_pool</entry> |
| <entry>String</entry> |
| <entry>(empty)</entry> |
| <entry> |
| Specifies the resource pool used by queries that insert into |
| the <codeph>sys.impala_query_log</codeph> table. If unset, no pool is applied. |
| </entry> |
| </row> |
| <row> |
| <entry>query_log_dml_exec_timeout_s</entry> |
| <entry>Integer</entry> |
| <entry>120</entry> |
| <entry> |
| Sets the <codeph>EXEC_TIME_LIMIT_S</codeph> query option |
| for the DML that inserts records into the <codeph>sys.impala_query_log</codeph> |
| table. |
| </entry> |
| </row> |
| </tbody> |
| </tgroup> |
| </table> |
| |
| </conbody> |
| </concept> |
| |
| <concept id="wm_observability"> |
| |
| <title>Observing</title> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="Observing"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| |
| <p> |
| The following metrics are available to help monitor and understand the behavior of |
| the Workload Management processing loop in Impala. |
| </p> |
| |
| <ul> |
| <li> |
| <codeph>impala-server.completed-queries.failure</codeph> - The number of |
| completed queries that failed to be written to the |
| <codeph>sys.impala_query_log</codeph> table. If the same query fails to be |
| written multiple times, each attempt increments this counter. |
| </li> |
| <li> |
| <codeph>impala-server.completed-queries.max-records-writes</codeph> - The number |
| of times completed queries were written to the <codeph>sys.impala_query_log</codeph> |
| table because the maximum records queued threshold was reached (set by the |
| <codeph>query_log_max_queued</codeph> startup flag). |
| </li> |
| <li> |
| <codeph>impala-server.completed-queries.queued</codeph> - The number of completed |
| queries that are currently queued and waiting to be written to the |
| <codeph>sys.impala_query_log</codeph> table. Matches the count of rows in |
| <codeph>sys.impala_query_live</codeph>. |
| </li> |
| <li> |
| <codeph>impala-server.completed-queries.scheduled-writes</codeph> - The number of |
| times completed queries were written to the <codeph>sys.impala_query_log</codeph> |
| table during the regularly scheduled interval. |
| </li> |
| <li> |
| <codeph>impala-server.completed-queries.write-durations</codeph> - A histogram |
| representing the time spent writing completed queries to the |
| <codeph>sys.impala_query_log</codeph> table. |
| </li> |
| <li> |
| <codeph>impala-server.completed-queries.written</codeph> - The number of |
| completed queries that were successfully written to the |
| <codeph>sys.impala_query_log</codeph> table. |
| </li> |
| </ul> |
| |
| <p> |
| The following log entries help track Workload Management processing activity. |
| </p> |
| |
| <table> |
| <tgroup cols="4" rowsep="1" colsep="1"> |
| <colspec colname="1" colwidth = "5*" /> |
| <colspec colname="1" colwidth = "3*" /> |
| <colspec colname="1" colwidth = "20*" /> |
| <colspec colname="1" colwidth = "20*" /> |
| <thead> |
| <row> |
| <entry>Daemon</entry> |
| <entry>Level</entry> |
| <entry>Description</entry> |
| <entry>Partial Log Text</entry> |
| </row> |
| </thead> |
| <tbody> |
| <row> |
| <entry>CatalogD</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Workload Management initialization has started, including create or upgrade of |
| <codeph>sys.impala_query_live</codeph> and <codeph>sys.impala_query_log</codeph> |
| tables. |
| </entry> |
| <entry> |
| <codeph>Starting workload management initialization</codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>CatalogD</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Workload Management initialization completed successfully. |
| </entry> |
| <entry> |
| <codeph>Completed workload management initialization</codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Shutdown of Workload Management has begun. Pending queries will be written |
| to <codeph>sys.impala_query_log</codeph> if possible. |
| </entry> |
| <entry> |
| <codeph>Workload management is shutting down</codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Workload Management shutdown completed. All pending queries written to |
| <codeph>sys.impala_query_log</codeph>. |
| </entry> |
| <entry> |
| <codeph>Workload management shutdown successful</codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Shutdown timed out before all queries were written to |
| <codeph>sys.impala_query_log</codeph>. The number of lost queries is logged. |
| </entry> |
| <entry> |
| <codeph> |
| Workload management shutdown timed out. Up to '5' queries may have been lost |
| </codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>ERROR</codeph></entry> |
| <entry> |
| The insert DML for moving completed queries to <codeph>sys.impala_query_log</codeph> |
| exceeded the maximum allowed SQL length. Lower the value of the |
| <codeph>query_log_max_queued</codeph> startup flag. |
| </entry> |
| <entry> |
| <codeph> |
| Completed queries table insert sql statement of length '9999999999' was |
| longer than the maximum of '2147483647', skipping |
| </codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>INFO</codeph></entry> |
| <entry> |
| Successfully inserted completed queries into |
| <codeph>sys.impala_query_log</codeph>. Log includes execution stats. |
| </entry> |
| <entry> |
| <codeph> |
| wrote completed queries table="sys.impala_query_log" record_count=1 |
| bytes="1.11 KB" gather_time="258.382us" exec_time="1s163ms" |
| query_id="9e4a49a89079233f:c72589eb00000000" |
| </codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>WARNING</codeph></entry> |
| <entry> |
| Failed to run the insert DML for moving completed queries telemetry to |
| <codeph>sys.impala_query_log</codeph>. Error details are in a subsequent log |
| message on the same coordinator. Ids of all completed queries who's telemetry |
| failed to be moved are logged in subsequent log messages at <codeph>ERROR</codeph> |
| level. |
| </entry> |
| <entry> |
| <codeph> |
| failed to write completed queries table="sys.impala_query_log" record_count=5 |
| bytes="1.11 KB" gather_time="258.382us" exec_time="1s163ms" |
| query_id="9e4a49a89079233f:c72589eb00000000" |
| </codeph> |
| </entry> |
| </row> |
| <row> |
| <entry>Coordinator</entry> |
| <entry><codeph>ERROR</codeph></entry> |
| <entry> |
| Indicates the telemetry for a single query failed to be moved to the |
| <codeph>sys.impala_query_log</codeph> table after three attempts. The telemetry |
| will be dropped and will no longer be available in either table. |
| </entry> |
| <entry> |
| <codeph> |
| could not write completed query table="sys.impala_query_log" |
| query_id="69477a18fdaadfd7:068f25bb00000000" |
| </codeph> |
| </entry> |
| </row> |
| </tbody> |
| </tgroup> |
| </table> |
| |
| </conbody> |
| </concept> |
| |
| <concept id="wm_maintenance"> |
| |
| <title>Table Maintenance</title> |
| |
| <prolog> |
| <metadata> |
| <data name="Category" value="Maintenance"/> |
| </metadata> |
| </prolog> |
| |
| <conbody> |
| |
| <ul> |
| <li> |
| <codeph>sys.impala_query_live</codeph> - No maintenance is required. This |
| table resides entirely in memory. |
| </li> |
| |
| <li> |
| <codeph>sys.impala_query_log</codeph> -- As an Iceberg table, it requires periodic |
| maintenance such as computing statistics and expiring old snapshots. Because |
| workloads vary, maintenance is not automated and must be performed manually. |
| <ul> |
| <li> |
| <codeph>OPTIMIZE TABLE sys.impala_query_log (FILE_SIZE_THRESHOLD_MB=128)</codeph> |
| </li> |
| <li> |
| <codeph>ALTER TABLE ice_tbl EXECUTE expire_snapshots(now() - interval 7 days);</codeph> |
| </li> |
| </ul> |
| </li> |
| </ul> |
| |
| </conbody> |
| </concept> |
| |
| </concept> |
