suneet-s commented on a change in pull request #10848:
URL: https://github.com/apache/druid/pull/10848#discussion_r580536570
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
Review comment:
```suggestion
Druid can store cache data in the local JVM heap or in an external
distributed key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
```
IMO It would be good to describe the defaults in this page. That way, as an
operator, this is the page that I can read to know what the initial state of
the system is and where to go to know how to change config options
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
-All caches have a pair of parameters that control the behavior of how
individual queries interact with the cache, a 'use' cache parameter, and a
'populate' cache parameter. These settings must be enabled at the service level
via [runtime properties](../configuration/index.md) to utilize cache, but can
be controlled on a per query basis by setting them on the [query
context](../querying/query-context.md). The 'use' parameter obviously controls
if a query will utilize cached results. The 'populate' parameter controls if a
query will update cached results. These are separate parameters to allow
queries on uncommon data to utilize cached results without polluting the cache
with results that are unlikely to be re-used by other queries, for example
large reports or very old data.
+### Segment caching
-## Query caching on Brokers
+The primary form of caching in Druid is the **segment cache**. The segment
cache stores query results on a per-segment basis. It is enabled on Historical
services by default.
-Brokers support both segment-level and whole-query result level caching.
Segment-level caching is controlled by the
-parameters `useCache` and `populateCache`. Whole-query result level caching is
controlled by the parameters
-`useResultLevelCache` and `populateResultLevelCache` and [runtime
properties](../configuration/index.md)
-`druid.broker.cache.*`.
+When your queries include data from segments that are mutable and undergoing
real-time ingestion, use a segment cache. In this case Druid caches query
results for immutable historical segments when possible. It re-computes results
for the real-time segments at query time.
-Enabling segment-level caching on the Broker can yield faster results than if
query caches were enabled on Historicals for small
-clusters. This is the recommended setup for smaller production clusters (< 5
servers). Populating segment-level caches on
-the Broker is _not_ recommended for large production clusters, since when the
property `druid.broker.cache.populateCache` is
-set to `true` (and query context parameter `populateCache` is _not_ set to
`false`), results from Historicals are returned
-on a per segment basis, and Historicals will not be able to do any local
result merging. This impairs the ability of the
-Druid cluster to scale well.
+For example, you have queries that frequently include incoming data from a
Kafka or Kinesis stream alongside unchanging segments. Whole-query caching is
not helpful in this scenario because the new data from real-time ingestion
continually invalidates the cache. Segment caching lets Druid cache results
from older immutable segments and merge them with updated data.
Review comment:
Does this belong under the next section?
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
Review comment:
Should this section include a description of "whole-query caching" as
well?
Maybe this is where we document that segment caching is enabled by default
while whole query caching is disabled
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
-All caches have a pair of parameters that control the behavior of how
individual queries interact with the cache, a 'use' cache parameter, and a
'populate' cache parameter. These settings must be enabled at the service level
via [runtime properties](../configuration/index.md) to utilize cache, but can
be controlled on a per query basis by setting them on the [query
context](../querying/query-context.md). The 'use' parameter obviously controls
if a query will utilize cached results. The 'populate' parameter controls if a
query will update cached results. These are separate parameters to allow
queries on uncommon data to utilize cached results without polluting the cache
with results that are unlikely to be re-used by other queries, for example
large reports or very old data.
+### Segment caching
-## Query caching on Brokers
+The primary form of caching in Druid is the **segment cache**. The segment
cache stores query results on a per-segment basis. It is enabled on Historical
services by default.
-Brokers support both segment-level and whole-query result level caching.
Segment-level caching is controlled by the
-parameters `useCache` and `populateCache`. Whole-query result level caching is
controlled by the parameters
-`useResultLevelCache` and `populateResultLevelCache` and [runtime
properties](../configuration/index.md)
-`druid.broker.cache.*`.
+When your queries include data from segments that are mutable and undergoing
real-time ingestion, use a segment cache. In this case Druid caches query
results for immutable historical segments when possible. It re-computes results
for the real-time segments at query time.
-Enabling segment-level caching on the Broker can yield faster results than if
query caches were enabled on Historicals for small
-clusters. This is the recommended setup for smaller production clusters (< 5
servers). Populating segment-level caches on
-the Broker is _not_ recommended for large production clusters, since when the
property `druid.broker.cache.populateCache` is
-set to `true` (and query context parameter `populateCache` is _not_ set to
`false`), results from Historicals are returned
-on a per segment basis, and Historicals will not be able to do any local
result merging. This impairs the ability of the
-Druid cluster to scale well.
+For example, you have queries that frequently include incoming data from a
Kafka or Kinesis stream alongside unchanging segments. Whole-query caching is
not helpful in this scenario because the new data from real-time ingestion
continually invalidates the cache. Segment caching lets Druid cache results
from older immutable segments and merge them with updated data.
-## Query caching on Historicals
+### Whole-query caching
-Historicals only support segment-level caching. Segment-level caching is
controlled by the query context
-parameters `useCache` and `populateCache` and [runtime
properties](../configuration/index.md)
-`druid.historical.cache.*`.
+If real-time ingestion invalidating the cache is not an issue for your
queries, you can use **whole-query caching** on the Broker to increase query
efficiency. Segment caching is less efficient in this case because it requires
Druid to merge the per-segment results for each query, even when the results
are cached. For instance, whole-query caching is a good option when you have
queries that include data from a batch ingestion task that runs every few hours
or once a day.
Review comment:
I think we should describe that whole query caching is done at the
broker, before fan out queries are sent to the historical. This is useful
because Druid will no longer need to merge the per-segment results back on the
broker.
Your description of the scenarios where this is helpful is great!
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
-All caches have a pair of parameters that control the behavior of how
individual queries interact with the cache, a 'use' cache parameter, and a
'populate' cache parameter. These settings must be enabled at the service level
via [runtime properties](../configuration/index.md) to utilize cache, but can
be controlled on a per query basis by setting them on the [query
context](../querying/query-context.md). The 'use' parameter obviously controls
if a query will utilize cached results. The 'populate' parameter controls if a
query will update cached results. These are separate parameters to allow
queries on uncommon data to utilize cached results without polluting the cache
with results that are unlikely to be re-used by other queries, for example
large reports or very old data.
+### Segment caching
-## Query caching on Brokers
+The primary form of caching in Druid is the **segment cache**. The segment
cache stores query results on a per-segment basis. It is enabled on Historical
services by default.
-Brokers support both segment-level and whole-query result level caching.
Segment-level caching is controlled by the
-parameters `useCache` and `populateCache`. Whole-query result level caching is
controlled by the parameters
-`useResultLevelCache` and `populateResultLevelCache` and [runtime
properties](../configuration/index.md)
-`druid.broker.cache.*`.
+When your queries include data from segments that are mutable and undergoing
real-time ingestion, use a segment cache. In this case Druid caches query
results for immutable historical segments when possible. It re-computes results
for the real-time segments at query time.
-Enabling segment-level caching on the Broker can yield faster results than if
query caches were enabled on Historicals for small
-clusters. This is the recommended setup for smaller production clusters (< 5
servers). Populating segment-level caches on
-the Broker is _not_ recommended for large production clusters, since when the
property `druid.broker.cache.populateCache` is
-set to `true` (and query context parameter `populateCache` is _not_ set to
`false`), results from Historicals are returned
-on a per segment basis, and Historicals will not be able to do any local
result merging. This impairs the ability of the
-Druid cluster to scale well.
+For example, you have queries that frequently include incoming data from a
Kafka or Kinesis stream alongside unchanging segments. Whole-query caching is
not helpful in this scenario because the new data from real-time ingestion
continually invalidates the cache. Segment caching lets Druid cache results
from older immutable segments and merge them with updated data.
-## Query caching on Historicals
+### Whole-query caching
-Historicals only support segment-level caching. Segment-level caching is
controlled by the query context
-parameters `useCache` and `populateCache` and [runtime
properties](../configuration/index.md)
-`druid.historical.cache.*`.
+If real-time ingestion invalidating the cache is not an issue for your
queries, you can use **whole-query caching** on the Broker to increase query
efficiency. Segment caching is less efficient in this case because it requires
Druid to merge the per-segment results for each query, even when the results
are cached. For instance, whole-query caching is a good option when you have
queries that include data from a batch ingestion task that runs every few hours
or once a day.
-Larger production clusters should enable segment-level cache population on
Historicals only (not on Brokers) to avoid
-having to use Brokers to merge all query results. Enabling cache population on
the Historicals instead of the Brokers
-enables the Historicals to do their own local result merging and puts less
strain on the Brokers.
+## Where to enable caching
-## Query caching on Ingestion Tasks
+**Segment cache** is available as follows:
+- On Brokers for small production clusters with less than five servers.
-Task executor processes such as the Peon or the experimental Indexer only
support segment-level caching. Segment-level
-caching is controlled by the query context parameters `useCache` and
`populateCache`
-and [runtime properties](../configuration/index.md) `druid.realtime.cache.*`.
+ Do not use segment caches on the Broker for large production clusters.
When `druid.broker.cache.populateCache` is `true` and query context parameter
`populateCache` _is not_ `false`, Historicals return results on a per-segment
basis without merging results locally thus negatively impacting cluster
scalability.
+
+- On Historicals, the default. Enable segment-level cache population on
Historicals for larger production clusters to prevent Brokers from having to
merge all query results. When you enable cache population on Historicals
instead of Brokers, the Historicals merge their own local results and put less
strain on the Brokers.
+
+- On ingestion tasks in the Peon or Indexer service. Larger production
clusters should enable segment-level cache population on task execution
services only to prevent Brokers from having to merge all query results. When
you enable cache population on task execution services instead of Brokers, the
the task execution services to merge their own local results and put less
strain on the Brokers.
+
+ Task executor services only support caches that store data locally. For
example the `caffeine` cache. This restriction exists because the cache stores
results at the level of intermediate partial segments generated by the
ingestion tasks. These intermediate partial segments may not be identical
across task replicas. Therefore task executor services ignore remote cache
types such as `memcached`.
+
+**Whole-query cache** is only available on Brokers.
+
+## Performance considerations for caching
+
+Consider the following as you develop your caching strategy:
+- Caching enables increased concurrency on the same system, therefore leading
to noticeable performance improvements for queries on Druid clusters handling
throughput for concurrent, mixed workloads.
+
+- If you are looking to improve response time for a single query or page load,
you should ignore caching. In general, response time for a single task should
meet performance objectives even when the cache is cold.
+
+- During query processing, the segment cache intercepts the query and sends
the results directly to the Broker. This way the query bypasses the data server
processing threads. For queries requiring minimal processing in the Broker,
cached queries are very quick. If work done on the Broker causes a query
bottleneck, enabling caching results in little noticeable query improvement.
+
+- The largest performance gains from segment caching tend to apply to `topN`
and time series queries. The impact is less for `groupBy` queries. The same
applies to queries with or without joins.
+
+### Scenarios where caching does not increase query performance
+
+Caching does not solve all types of query peformance issues. For each cache
type there are scenarios where caching is likely to be of little benefit.
+
+**Segment caching** doesn't work for TBD.
+
+**Whole-query caching** doesn't work for the following:
+- `GroupBy` v2 queries
+- queries with joins
+- queries on real-time data.
Review comment:
I think we shouldn't mention this here. The other 2 in the list are
specific limitations of the Druid system. This point talks about the fact that
if you are querying real time data (last 5 minutes), you will never have a
cached response since data for the last 5 minutes is always coming in. What do
you think?
```suggestion
```
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
-All caches have a pair of parameters that control the behavior of how
individual queries interact with the cache, a 'use' cache parameter, and a
'populate' cache parameter. These settings must be enabled at the service level
via [runtime properties](../configuration/index.md) to utilize cache, but can
be controlled on a per query basis by setting them on the [query
context](../querying/query-context.md). The 'use' parameter obviously controls
if a query will utilize cached results. The 'populate' parameter controls if a
query will update cached results. These are separate parameters to allow
queries on uncommon data to utilize cached results without polluting the cache
with results that are unlikely to be re-used by other queries, for example
large reports or very old data.
+### Segment caching
-## Query caching on Brokers
+The primary form of caching in Druid is the **segment cache**. The segment
cache stores query results on a per-segment basis. It is enabled on Historical
services by default.
-Brokers support both segment-level and whole-query result level caching.
Segment-level caching is controlled by the
-parameters `useCache` and `populateCache`. Whole-query result level caching is
controlled by the parameters
-`useResultLevelCache` and `populateResultLevelCache` and [runtime
properties](../configuration/index.md)
-`druid.broker.cache.*`.
+When your queries include data from segments that are mutable and undergoing
real-time ingestion, use a segment cache. In this case Druid caches query
results for immutable historical segments when possible. It re-computes results
for the real-time segments at query time.
-Enabling segment-level caching on the Broker can yield faster results than if
query caches were enabled on Historicals for small
-clusters. This is the recommended setup for smaller production clusters (< 5
servers). Populating segment-level caches on
-the Broker is _not_ recommended for large production clusters, since when the
property `druid.broker.cache.populateCache` is
-set to `true` (and query context parameter `populateCache` is _not_ set to
`false`), results from Historicals are returned
-on a per segment basis, and Historicals will not be able to do any local
result merging. This impairs the ability of the
-Druid cluster to scale well.
+For example, you have queries that frequently include incoming data from a
Kafka or Kinesis stream alongside unchanging segments. Whole-query caching is
not helpful in this scenario because the new data from real-time ingestion
continually invalidates the cache. Segment caching lets Druid cache results
from older immutable segments and merge them with updated data.
-## Query caching on Historicals
+### Whole-query caching
-Historicals only support segment-level caching. Segment-level caching is
controlled by the query context
-parameters `useCache` and `populateCache` and [runtime
properties](../configuration/index.md)
-`druid.historical.cache.*`.
+If real-time ingestion invalidating the cache is not an issue for your
queries, you can use **whole-query caching** on the Broker to increase query
efficiency. Segment caching is less efficient in this case because it requires
Druid to merge the per-segment results for each query, even when the results
are cached. For instance, whole-query caching is a good option when you have
queries that include data from a batch ingestion task that runs every few hours
or once a day.
-Larger production clusters should enable segment-level cache population on
Historicals only (not on Brokers) to avoid
-having to use Brokers to merge all query results. Enabling cache population on
the Historicals instead of the Brokers
-enables the Historicals to do their own local result merging and puts less
strain on the Brokers.
+## Where to enable caching
-## Query caching on Ingestion Tasks
+**Segment cache** is available as follows:
+- On Brokers for small production clusters with less than five servers.
-Task executor processes such as the Peon or the experimental Indexer only
support segment-level caching. Segment-level
-caching is controlled by the query context parameters `useCache` and
`populateCache`
-and [runtime properties](../configuration/index.md) `druid.realtime.cache.*`.
+ Do not use segment caches on the Broker for large production clusters.
When `druid.broker.cache.populateCache` is `true` and query context parameter
`populateCache` _is not_ `false`, Historicals return results on a per-segment
basis without merging results locally thus negatively impacting cluster
scalability.
+
+- On Historicals, the default. Enable segment-level cache population on
Historicals for larger production clusters to prevent Brokers from having to
merge all query results. When you enable cache population on Historicals
instead of Brokers, the Historicals merge their own local results and put less
strain on the Brokers.
+
+- On ingestion tasks in the Peon or Indexer service. Larger production
clusters should enable segment-level cache population on task execution
services only to prevent Brokers from having to merge all query results. When
you enable cache population on task execution services instead of Brokers, the
the task execution services to merge their own local results and put less
strain on the Brokers.
+
+ Task executor services only support caches that store data locally. For
example the `caffeine` cache. This restriction exists because the cache stores
results at the level of intermediate partial segments generated by the
ingestion tasks. These intermediate partial segments may not be identical
across task replicas. Therefore task executor services ignore remote cache
types such as `memcached`.
+
+**Whole-query cache** is only available on Brokers.
+
+## Performance considerations for caching
+
+Consider the following as you develop your caching strategy:
+- Caching enables increased concurrency on the same system, therefore leading
to noticeable performance improvements for queries on Druid clusters handling
throughput for concurrent, mixed workloads.
+
+- If you are looking to improve response time for a single query or page load,
you should ignore caching. In general, response time for a single task should
meet performance objectives even when the cache is cold.
+
+- During query processing, the segment cache intercepts the query and sends
the results directly to the Broker. This way the query bypasses the data server
processing threads. For queries requiring minimal processing in the Broker,
cached queries are very quick. If work done on the Broker causes a query
bottleneck, enabling caching results in little noticeable query improvement.
+
+- The largest performance gains from segment caching tend to apply to `topN`
and time series queries. The impact is less for `groupBy` queries. The same
applies to queries with or without joins.
+
+### Scenarios where caching does not increase query performance
+
+Caching does not solve all types of query peformance issues. For each cache
type there are scenarios where caching is likely to be of little benefit.
+
+**Segment caching** doesn't work for TBD.
Review comment:
@abhishekagarwal87 You did some investigation into segment caching a
little while ago. Do you happen to know where to find the list of places where
segment caching isn't supported yet
##########
File path: docs/querying/using-caching.md
##########
@@ -0,0 +1,92 @@
+---
+id: using-caching
+title: "Using query caching"
+---
+
+<!--
+ ~ 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.
+ -->
+
+
+This topic covers how to configure services to populate and use the Druid
query cache. For a conceptual overview and use cases, see [Query
caching](./caching.md). For information on how to configure the caching
mechanism, see [Cache
configuration](../configuration/index.md#cache-configuration).
+
+All query caches have a pair of parameters that control the way individual
queries interact with the cache:
+
+- `useCache` to instruct queries to use the cache for results.
+- `populateCache` to instruct a query to cache its results.
+
+The separation of concerns, usage and population, lets you include cached
results for queries on uncommon data without polluting the cache with results
that are unlikely to be reused by other queries, for example, large reports or
queries on very old data.
+
+To use caching, you must first enable the settings for a service in the
runtime properties. Afterward, you can control which queries use the cache or
populate the cache on a per-query basis within the query context.
+
+The query context settings do not override the server settings. For example,
it is possible to set a query context to use the cache, but if it is not
enabled at the service level, there's no cache to use.
+
+## Enabling query caching on Brokers
+Brokers support both segment-level and whole-query result level caching.
+
+To control **segment caching** on the Broker, set the `useCache` and
`populateCache`runtime properties. For example, to set the Broker to use and
populate the segment cache for queries:
+```
+druid.broker.cache.useCache=true
+druid.broker.cache.populateCache=true
+```
+To control **whole-query caching** on the Broker, set the
`useResultLevelCache` and `populateResultLevelCache` runtime properties. For
example, to set the Broker to use and populate the whole-query cache for
queries:
+```
+druid.broker.cache.useResultLevelCache=true
+druid.broker.cache.populateResultLevelCache=true
+```
+See [Broker caching](../configuration/index.md#broker-caching) for a
description of all available Broker cache configurations.
+
+## Enabling query caching on Historicals
+Historicals only support **segment-level** caching. To control caching on the
Historical, set the `useCache` and `populateCache` runtime properties. For
example, to set the Historical to both use and populate the segment cache for
queries:
+ ```
+ druid.broker.cache.useCache=true
+ druid.broker.cache.populateCache=true
+ ```
+See [Historical caching](../configuration/index.md#historical-caching) for a
description of all available Historical cache configurations.
+
+## Enabling query caching on task executor services
+Task executor services, the Peon or the Indexer, only support
**segment-level** caching. To control caching on a task executor service, set
the `useCache` and `populateCache` runtime properties. For example, to set the
Peon to both use and populate the segment cache for queries:
+
+```
+druid.realtime.cache.useCache=true
+druid.realtime.cache.populateCache=true
+```
+
+See [Peon caching](configuration/index.md#peon-caching) for a description of
all available task executor service caching options.
+
+## Enabling caching in the query context
+After you enable caching for a service, set cache options for individual
queries in the query [context](./query-context.md). For example, you can `POST`
a Druid SQL request to the HTTP POST API and include the context as a JSON
object:
+
+```
+{
+ "query" : "SELECT COUNT(*) FROM data_source WHERE foo = 'bar' AND __time >
TIMESTAMP '2020-01-01 00:00:00'",
+ "context" : {
+ "useCache" : "true",
+ "populateCache" : "false"
+ }
+}
+```
+In this example `populateCache` is `false` because the query references data
that is over a year old. For more information, see [Druid SQL client
APIs](./sql.md#client-apis).
+
+reiterate that populateCache : true must be set somehwere on at least 1
service. By defaults it's on on Historical and that serves the majority of use
cases.
Review comment:
needs a touch up
##########
File path: docs/querying/using-caching.md
##########
@@ -0,0 +1,92 @@
+---
+id: using-caching
+title: "Using query caching"
+---
+
+<!--
+ ~ 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.
+ -->
+
+
+This topic covers how to configure services to populate and use the Druid
query cache. For a conceptual overview and use cases, see [Query
caching](./caching.md). For information on how to configure the caching
mechanism, see [Cache
configuration](../configuration/index.md#cache-configuration).
+
+All query caches have a pair of parameters that control the way individual
queries interact with the cache:
+
+- `useCache` to instruct queries to use the cache for results.
+- `populateCache` to instruct a query to cache its results.
+
+The separation of concerns, usage and population, lets you include cached
results for queries on uncommon data without polluting the cache with results
that are unlikely to be reused by other queries, for example, large reports or
queries on very old data.
+
+To use caching, you must first enable the settings for a service in the
runtime properties. Afterward, you can control which queries use the cache or
populate the cache on a per-query basis within the query context.
+
+The query context settings do not override the server settings. For example,
it is possible to set a query context to use the cache, but if it is not
enabled at the service level, there's no cache to use.
+
+## Enabling query caching on Brokers
+Brokers support both segment-level and whole-query result level caching.
+
+To control **segment caching** on the Broker, set the `useCache` and
`populateCache`runtime properties. For example, to set the Broker to use and
populate the segment cache for queries:
+```
+druid.broker.cache.useCache=true
+druid.broker.cache.populateCache=true
+```
+To control **whole-query caching** on the Broker, set the
`useResultLevelCache` and `populateResultLevelCache` runtime properties. For
example, to set the Broker to use and populate the whole-query cache for
queries:
+```
+druid.broker.cache.useResultLevelCache=true
+druid.broker.cache.populateResultLevelCache=true
+```
+See [Broker caching](../configuration/index.md#broker-caching) for a
description of all available Broker cache configurations.
+
+## Enabling query caching on Historicals
+Historicals only support **segment-level** caching. To control caching on the
Historical, set the `useCache` and `populateCache` runtime properties. For
example, to set the Historical to both use and populate the segment cache for
queries:
+ ```
+ druid.broker.cache.useCache=true
+ druid.broker.cache.populateCache=true
+ ```
+See [Historical caching](../configuration/index.md#historical-caching) for a
description of all available Historical cache configurations.
+
+## Enabling query caching on task executor services
+Task executor services, the Peon or the Indexer, only support
**segment-level** caching. To control caching on a task executor service, set
the `useCache` and `populateCache` runtime properties. For example, to set the
Peon to both use and populate the segment cache for queries:
+
+```
+druid.realtime.cache.useCache=true
+druid.realtime.cache.populateCache=true
+```
+
+See [Peon caching](configuration/index.md#peon-caching) for a description of
all available task executor service caching options.
+
+## Enabling caching in the query context
+After you enable caching for a service, set cache options for individual
queries in the query [context](./query-context.md). For example, you can `POST`
a Druid SQL request to the HTTP POST API and include the context as a JSON
object:
+
+```
+{
+ "query" : "SELECT COUNT(*) FROM data_source WHERE foo = 'bar' AND __time >
TIMESTAMP '2020-01-01 00:00:00'",
+ "context" : {
+ "useCache" : "true",
+ "populateCache" : "false"
+ }
+}
+```
+In this example `populateCache` is `false` because the query references data
that is over a year old. For more information, see [Druid SQL client
APIs](./sql.md#client-apis).
Review comment:
`because the query references data that is over a year old. `
This reads as if the system is smart enough to make this call. I think we
want to say something along the lines of the user making the query does not
want to fill up the cache with results for old segments, so they may decide to
disable populating the cache for old data, but still want to use the cache for
recent data. In this scenario, they'd set the query context like this. What do
you think?
##########
File path: docs/querying/caching.md
##########
@@ -22,63 +22,75 @@ title: "Query caching"
~ under the License.
-->
+You can enable caching in Apache Druid to improve query times for frequently
accessed data. This topic defines the different types of caching for Druid. It
describes the default caching behavior and provides guidance and examples to
help you hone your caching strategy.
-Apache Druid supports query result caching at both the segment and whole-query
result level. Cache data can be stored in the
-local JVM heap or in an external distributed key/value store. In all cases,
the Druid cache is a query result cache.
-The only difference is whether the result is a _partial result_ for a
particular segment, or the result for an entire
-query. In both cases, the cache is invalidated as soon as any underlying data
changes; it will never return a stale
-result.
+If you're unfamiliar with Druid architecture, review the following topics
before proceeding with caching:
+- [Druid Design](../design/architecture.md)
+- [Segments](../design/segments.md)
+- [Query execution](./query-execution)
-Segment-level caching allows the cache to be leveraged even when some of the
underling segments are mutable and
-undergoing real-time ingestion. In this case, Druid will potentially cache
query results for immutable historical
-segments, while re-computing results for the real-time segments on each query.
Whole-query result level caching is not
-useful in this scenario, since it would be continuously invalidated.
+For instructions to configure caching see [Using query
caching](./using-caching.md).
-Segment-level caching does require Druid to merge the per-segment results on
each query, even when they are served
-from the cache. For this reason, whole-query result level caching can be more
efficient if invalidation due to real-time
-ingestion is not an issue.
+## Cache types
+Druid supports segment caching which stores _partial results_ of a query for a
specific segment, and whole-query caching which stores all results for a query.
To avoid returning stale results, Druid invalidates the cache the moment any
underlying data changes for both types of cache.
-## Using and populating cache
+Druid can store cache data the local JVM heap or in an external distributed
key/value store. See [Cache
configuration](../configuration/index.md#cache-configuration) for information
on how to configure cache storage.
-All caches have a pair of parameters that control the behavior of how
individual queries interact with the cache, a 'use' cache parameter, and a
'populate' cache parameter. These settings must be enabled at the service level
via [runtime properties](../configuration/index.md) to utilize cache, but can
be controlled on a per query basis by setting them on the [query
context](../querying/query-context.md). The 'use' parameter obviously controls
if a query will utilize cached results. The 'populate' parameter controls if a
query will update cached results. These are separate parameters to allow
queries on uncommon data to utilize cached results without polluting the cache
with results that are unlikely to be re-used by other queries, for example
large reports or very old data.
+### Segment caching
-## Query caching on Brokers
+The primary form of caching in Druid is the **segment cache**. The segment
cache stores query results on a per-segment basis. It is enabled on Historical
services by default.
-Brokers support both segment-level and whole-query result level caching.
Segment-level caching is controlled by the
-parameters `useCache` and `populateCache`. Whole-query result level caching is
controlled by the parameters
-`useResultLevelCache` and `populateResultLevelCache` and [runtime
properties](../configuration/index.md)
-`druid.broker.cache.*`.
+When your queries include data from segments that are mutable and undergoing
real-time ingestion, use a segment cache. In this case Druid caches query
results for immutable historical segments when possible. It re-computes results
for the real-time segments at query time.
-Enabling segment-level caching on the Broker can yield faster results than if
query caches were enabled on Historicals for small
-clusters. This is the recommended setup for smaller production clusters (< 5
servers). Populating segment-level caches on
-the Broker is _not_ recommended for large production clusters, since when the
property `druid.broker.cache.populateCache` is
-set to `true` (and query context parameter `populateCache` is _not_ set to
`false`), results from Historicals are returned
-on a per segment basis, and Historicals will not be able to do any local
result merging. This impairs the ability of the
-Druid cluster to scale well.
+For example, you have queries that frequently include incoming data from a
Kafka or Kinesis stream alongside unchanging segments. Whole-query caching is
not helpful in this scenario because the new data from real-time ingestion
continually invalidates the cache. Segment caching lets Druid cache results
from older immutable segments and merge them with updated data.
-## Query caching on Historicals
+### Whole-query caching
-Historicals only support segment-level caching. Segment-level caching is
controlled by the query context
-parameters `useCache` and `populateCache` and [runtime
properties](../configuration/index.md)
-`druid.historical.cache.*`.
+If real-time ingestion invalidating the cache is not an issue for your
queries, you can use **whole-query caching** on the Broker to increase query
efficiency. Segment caching is less efficient in this case because it requires
Druid to merge the per-segment results for each query, even when the results
are cached. For instance, whole-query caching is a good option when you have
queries that include data from a batch ingestion task that runs every few hours
or once a day.
-Larger production clusters should enable segment-level cache population on
Historicals only (not on Brokers) to avoid
-having to use Brokers to merge all query results. Enabling cache population on
the Historicals instead of the Brokers
-enables the Historicals to do their own local result merging and puts less
strain on the Brokers.
+## Where to enable caching
-## Query caching on Ingestion Tasks
+**Segment cache** is available as follows:
+- On Brokers for small production clusters with less than five servers.
Review comment:
nit: Maybe this should be last in the list, since it's the least
impactful.
Caching on the historicals - applies to almost everyone.
Caching on the ingestion tasks - useful if you're querying realtime data
Caching on the broker - not useful in production scenarios... I can't think
of a reason to ever to this tbh
##########
File path: docs/querying/using-caching.md
##########
@@ -0,0 +1,92 @@
+---
+id: using-caching
+title: "Using query caching"
+---
+
+<!--
+ ~ 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.
+ -->
+
+
+This topic covers how to configure services to populate and use the Druid
query cache. For a conceptual overview and use cases, see [Query
caching](./caching.md). For information on how to configure the caching
mechanism, see [Cache
configuration](../configuration/index.md#cache-configuration).
+
+All query caches have a pair of parameters that control the way individual
queries interact with the cache:
+
+- `useCache` to instruct queries to use the cache for results.
+- `populateCache` to instruct a query to cache its results.
+
+The separation of concerns, usage and population, lets you include cached
results for queries on uncommon data without polluting the cache with results
that are unlikely to be reused by other queries, for example, large reports or
queries on very old data.
+
+To use caching, you must first enable the settings for a service in the
runtime properties. Afterward, you can control which queries use the cache or
populate the cache on a per-query basis within the query context.
+
+The query context settings do not override the server settings. For example,
it is possible to set a query context to use the cache, but if it is not
enabled at the service level, there's no cache to use.
Review comment:
I think this is a little more nuanced. It doesn't enable the feature if
it's disabled in the server settings, but a query can specify that it doesn't
want to populate the cache or use the cache, even if those settings are made in
the server settings
----------------------------------------------------------------
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.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
