techdocsmith commented on a change in pull request #12128:
URL: https://github.com/apache/druid/pull/12128#discussion_r780446061



##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.

Review comment:
       I don't know if this is in the style, but recommend "Apache Druid" at 
first mention.
   
   I wonder if we can change the setup of the introduction a little bit to 
answer:
   - what are concurrent, mixed (heterogenerous) workloads? Can we provide an 
example?
   - what are the problems you can run into if you don't optimize (like 
long-running lower-priority  queries will impact the perf of higher priority 
queries)?
   Then talk about how this topic talks about strategies to optimize 
Druid/avoid perf problems.

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.

Review comment:
       I like the freeway lanes analogy!
   
   Something about "low" lanes and "high" lanes feels undefined to me. Perhaps 
a sentence stating that you define the high-priority query lanes and 
low-priority query lanes?
   
   Also wonder if we can use an example here. Maybe if you set up an example in 
the intro, you can just refer to the same high priority/low priority example 
throughout?

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.
+
+There are three types of load rules: forever, interval, and period. Select the 
load rule that matches your use case for each Historical, whether you want all 
segments to be loaded, segments within a certain time interval, or segments 
within a certain time period. 
+
+In the load rule, define tiers in the `tieredReplicants` property. Provide 
descriptive names for your tiers, and specify how many replicas each tier 
should have. You can designate a higher number of replicas for the hot tier to 
increase the concurrency for processing queries.
+
+Example load rule with two Historical tiers, named “hot” and “\_default\_tier”:
+
+```
+{
+  "type" : "loadByPeriod",
+  "period" : "P1M",
+  "includeFuture" : true,
+  "tieredReplicants": {
+    "hot": 3,
+    "_default_tier" : 1
+  }
+}
+```
+
+See [Load rules](rule-configuration.md#load-rules) for more information on 
segment load rules. Visit [Tutorial: Configuring data 
retention](../tutorials/tutorial-retention.md) for an example of setting 
retention rules from the Druid web console.
+
+### Assign Historicals to tiers
+
+Assign the Historical to tiers by labeling the tier name and setting the 
priority value in the  `historical/runtime.properties` files.
+
+Example config for a Historical in the hot tier:
+
+```
+druid.server.tier=hot
+druid.server.priority=1
+```
+
+Example config for a Historical in the cold tier:
+
+```
+druid.server.tier=_default_tier
+druid.server.priority=0
+```
+
+See [Historical general 
configuration](../configuration/index.md#historical-general-configuration) for 
more details on these properties.
+
+## Broker tiering
+
+To set up Broker tiering, assign Brokers to tiers and configure query routing 
by the Router. You must set up Historical tiering before you can use Broker 
tiering.

Review comment:
       Why would I do Broker tiering? What is the added benefit. Consider 
stating that it's dependent on Historical tiering first.

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.
+
+There are three types of load rules: forever, interval, and period. Select the 
load rule that matches your use case for each Historical, whether you want all 
segments to be loaded, segments within a certain time interval, or segments 
within a certain time period. 
+
+In the load rule, define tiers in the `tieredReplicants` property. Provide 
descriptive names for your tiers, and specify how many replicas each tier 
should have. You can designate a higher number of replicas for the hot tier to 
increase the concurrency for processing queries.
+
+Example load rule with two Historical tiers, named “hot” and “\_default\_tier”:
+
+```
+{
+  "type" : "loadByPeriod",
+  "period" : "P1M",
+  "includeFuture" : true,
+  "tieredReplicants": {
+    "hot": 3,
+    "_default_tier" : 1
+  }
+}
+```
+
+See [Load rules](rule-configuration.md#load-rules) for more information on 
segment load rules. Visit [Tutorial: Configuring data 
retention](../tutorials/tutorial-retention.md) for an example of setting 
retention rules from the Druid web console.
+
+### Assign Historicals to tiers
+
+Assign the Historical to tiers by labeling the tier name and setting the 
priority value in the  `historical/runtime.properties` files.
+
+Example config for a Historical in the hot tier:

Review comment:
       ```suggestion
   Example Historical in the hot tier:
   ```

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.
+
+There are three types of load rules: forever, interval, and period. Select the 
load rule that matches your use case for each Historical, whether you want all 
segments to be loaded, segments within a certain time interval, or segments 
within a certain time period. 
+
+In the load rule, define tiers in the `tieredReplicants` property. Provide 
descriptive names for your tiers, and specify how many replicas each tier 
should have. You can designate a higher number of replicas for the hot tier to 
increase the concurrency for processing queries.
+
+Example load rule with two Historical tiers, named “hot” and “\_default\_tier”:
+
+```
+{
+  "type" : "loadByPeriod",
+  "period" : "P1M",
+  "includeFuture" : true,
+  "tieredReplicants": {
+    "hot": 3,
+    "_default_tier" : 1
+  }
+}
+```
+
+See [Load rules](rule-configuration.md#load-rules) for more information on 
segment load rules. Visit [Tutorial: Configuring data 
retention](../tutorials/tutorial-retention.md) for an example of setting 
retention rules from the Druid web console.
+
+### Assign Historicals to tiers
+
+Assign the Historical to tiers by labeling the tier name and setting the 
priority value in the  `historical/runtime.properties` files.
+
+Example config for a Historical in the hot tier:
+
+```
+druid.server.tier=hot
+druid.server.priority=1
+```
+
+Example config for a Historical in the cold tier:

Review comment:
       ```suggestion
   Example Historical in the cold tier:
   ```

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.

Review comment:
       The blunt statement about how Druid stores data, feels abrupt. It's also 
not exactly the subject of this article. Maybe talk about it in the context of 
data distribution across Historicals.
   
   Also is this date-related tiering? it seems like the statement  in line 78 
about fast-running queries vs slow-running queries doesn't apply here? Like we 
could issue a really slow query on the most current data, right?
   
   
   

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.

Review comment:
       - I'm missing the conceptual introduction like was there for Query 
laning.
   - Are Historical tiering and Broker tiering coupled in a single approach? If 
so, I think they should go into a "Service tiering" or "Historical and Broker 
service tiering" section.
   - Avoid "we'll" You can configure Brokers... 
   - What are "hot" queries? "cold" queries?

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.
+
+There are three types of load rules: forever, interval, and period. Select the 
load rule that matches your use case for each Historical, whether you want all 
segments to be loaded, segments within a certain time interval, or segments 
within a certain time period. 
+
+In the load rule, define tiers in the `tieredReplicants` property. Provide 
descriptive names for your tiers, and specify how many replicas each tier 
should have. You can designate a higher number of replicas for the hot tier to 
increase the concurrency for processing queries.
+
+Example load rule with two Historical tiers, named “hot” and “\_default\_tier”:

Review comment:
       make sure to include a description of what it does (loads only the last 
month of data on these tiers?)

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.

Review comment:
       ```suggestion
   Druid provides the following approaches to isolate resources and improve 
query concurrency:
   - **Query laning** where you set a limit on the maximum number of 
long-running queries executed on each Broker. 
   - **Cluster tiering** which defines separate groups of Historicals and 
Brokers to receive different query assignments based on query priority.
   ```
   Consider giving the options more space and make them easier to scan. Avoid 
naming a specific number of options when possible, to make it simpler to add 
new options.

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 

Review comment:
       Even though Broker tiering is optional, it is dependent on Historical 
Tiering, so I'd keep them in the same section. In the beginning I'd make it 
clear why you'd want to choose to do tiering for Historicals only or both 
Historicals and Brokers.

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.

Review comment:
       If Broker tiering is the ideal solution, should we put it first as the 
option. @jihoonson how can we help the user decide when to choose one approach 
over the other? For example, why would you choose Query laning if Broker 
tiering is ideal?
   
   Does it make sense to have a Stratgies section like this one: 
https://druid.apache.org/docs/0.22.1/ingestion/compaction.html#compaction-strategies

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.
+
+In Druid, query lanes reserve resources for Broker HTTP threads. Each Druid 
query requires one Broker thread. The number of threads on a Broker is defined 
by the `druid.server.http.numThreads` parameter. Broker threads may be occupied 
by tasks other than queries, such as health checks. You can use query laning to 
limit the number of HTTP threads designated for resource-intensive queries, 
leaving other threads available for short-running queries and other tasks.
+
+### General properties
+
+Set the following query laning properties in the `broker/runtime.properties` 
file.
+
+* `druid.query.scheduler.numThreads` – The total number of queries that can be 
served per Broker. We recommend setting this value to 1-2 less than 
`druid.server.http.numThreads`.
+  > The query scheduler by default does not limit the number of Broker HTTP 
threads. Setting this property to a bounded number limits the thread count. If 
the allocated threads are all occupied, any incoming query, including 
interactive queries, will be rejected with an HTTP 429 status code.
+
+* `druid.query.scheduler.laning.strategy` – The strategy used to assign 
queries to lanes. You can [define your own laning strategy 
manually](../configuration/index.md#manual-laning-strategy) or use the built-in 
[“high/low” laning strategy](../configuration/index.md#highlow-laning-strategy).
+
+Consider also defining a [prioritization 
strategy](../configuration/index.md#prioritization-strategies) for how queries 
are labeled high or low priority. Otherwise, manually set the priority for 
incoming queries on the query context.
+
+### Lane-specific properties
+
+If you use a manual laning strategy, set the following:
+
+* `druid.query.scheduler.laning.lanes.{name}`
+* `druid.query.scheduler.laning.isLimitPercent`
+
+If you use the high/low laning strategy, set the following:
+
+* `druid.query.scheduler.laning.maxLowPercent` – The maximum percent of query 
threads to handle low priority queries. The remaining query threads are 
dedicated to higher priority queries.
+
+### Example
+
+Example config for query laning with the high/low laning strategy:
+
+```
+# Limit the number of HTTP threads for query processing
+# This value should be less than druid.server.http.numThreads
+druid.query.scheduler.numThreads=40 
+
+# Laning strategy
+druid.query.scheduler.laning.strategy=hilo
+druid.query.scheduler.laning.maxLowPercent=20
+```
+
+See [Query prioritization and 
laning](../configuration/index.md#query-prioritization-and-laning) for details 
on query laning and the available query laning strategies.
+
+## Historical tiering
+
+Configure segment loading and tiers for Historicals as described in this 
section. In the examples below, we set up two tiers—hot and cold—for the 
Historicals and for the Brokers. We’ll instruct the Brokers to serve hot 
queries before cold queries. Short-running queries will be routed to the hot 
tiers, and long-running queries will be routed to the cold tiers.
+
+It is possible to separate Historical processes into tiers without having 
separate Broker tiers. This way, you can assign data from specific time 
intervals to specific tiers in order to support higher concurrency on hot data. 
+
+### Configure segment loading
+
+Druid stores data in segment files. Define a [load 
rule](rule-configuration.md#load-rules) to indicate how segment replicas should 
be assigned to different Historical tiers. For example, you may store segments 
of more recent data on more powerful hardware for better performance.
+
+There are three types of load rules: forever, interval, and period. Select the 
load rule that matches your use case for each Historical, whether you want all 
segments to be loaded, segments within a certain time interval, or segments 
within a certain time period. 
+
+In the load rule, define tiers in the `tieredReplicants` property. Provide 
descriptive names for your tiers, and specify how many replicas each tier 
should have. You can designate a higher number of replicas for the hot tier to 
increase the concurrency for processing queries.
+
+Example load rule with two Historical tiers, named “hot” and “\_default\_tier”:
+
+```
+{
+  "type" : "loadByPeriod",
+  "period" : "P1M",
+  "includeFuture" : true,
+  "tieredReplicants": {
+    "hot": 3,
+    "_default_tier" : 1
+  }
+}
+```
+
+See [Load rules](rule-configuration.md#load-rules) for more information on 
segment load rules. Visit [Tutorial: Configuring data 
retention](../tutorials/tutorial-retention.md) for an example of setting 
retention rules from the Druid web console.
+
+### Assign Historicals to tiers
+
+Assign the Historical to tiers by labeling the tier name and setting the 
priority value in the  `historical/runtime.properties` files.

Review comment:
       ```suggestion
   To assign a Historical to a tier, add a label for the tier name and set the 
priority value in the  `historical/runtime.properties` for the Historical.
   ```

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency
+sidebar_label: Query concurrency
+---
+
+<!--
+  ~ 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.
+  -->
+
+If you frequently run concurrent, mixed workloads on your Druid cluster, 
configure Druid to properly allocate cluster resources and optimize your 
overall query performance. With proper resource isolation, you can execute 
long-running, low priority queries that are resource intensive without 
interfering with short-running, high priority queries that require fewer 
resources. By separating cluster resources, you prevent queries from competing 
with each other for resources such as CPU, memory, and network access.
+
+There are two approaches to isolate your resources for improving query 
concurrency: query laning and cluster tiering. Use query laning to set a limit 
on the maximum number of long-running queries executed on each Broker. Use 
cluster tiering to define separate groups of Historicals and Brokers to which 
different queries can be directed based on their priority.
+
+## Query laning
+
+Query laning directs Druid to restrict resource usage for less urgent queries 
to ensure dedicated resources for higher priority queries. Query laning is 
ideal when you need to run many concurrent queries having heterogeneous 
workloads.
+
+Query lanes are analogous to carpool and normal lanes on the freeway. With 
query laning, Druid restricts low priority queries to low lanes and allows high 
priority queries to run wherever possible, whether in a high or low lane. In 
this way, higher priority queries may bypass other queries in lower priority 
lanes.

Review comment:
       The piece I am missing from this topic is how a Broker (or Druid) 
decides which query goes into which lane. I think it merits a brief mention of 
how you can specify a lane in the query context or set up thresholds 
(depinding). Then, in your example, show a query getting routed to high/low 
lanes.

##########
File path: docs/operations/query-concurrency.md
##########
@@ -0,0 +1,174 @@
+---
+id: query-concurrency
+title: Query concurrency

Review comment:
       I agree with @jihoonson . Configure Druid for mixed use workloads (this 
title would be very long for the left nav, but you get the idea). This is 
mostly task based, so a "verb" title would be good.




-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to