Fast, simple, reliable. HikariCP is a "zero-overhead" production ready JDBC connection pool. At roughly 130Kb, the library is very light. Read about how we do it here.
"Simplicity is prerequisite for reliability."
- Edsger Dijkstra
Java 8 maven artifact:
<dependency>
<groupId>com.zaxxer</groupId>
<artifactId>HikariCP</artifactId>
<version>2.6.1</version>
</dependency>
Java 7 maven artifact (maintenance mode):
<dependency>
<groupId>com.zaxxer</groupId>
<artifactId>HikariCP-java7</artifactId>
<version>2.4.11</version>
</dependency>
Java 6 maven artifact (maintenance mode):
<dependency>
<groupId>com.zaxxer</groupId>
<artifactId>HikariCP-java6</artifactId>
<version>2.3.13</version>
</dependency>
Microbenchmarks were created to isolate and measure the overhead of pools using the JMH microbenchmark framework. You can checkout the HikariCP benchmark project for details and review/run the benchmarks yourself.
- One Connection Cycle is defined as single
DataSource.getConnection()
/Connection.close()
. - One Statement Cycle is defined as single
Connection.prepareStatement()
,Statement.execute()
,Statement.close()
.
2 Intel Core i7-3770 CPU @ 3.40GHz
3 Uncontended benchmark: 32 threads/32 connections, Contended benchmark: 32 threads, 16 connections
4 Apache Tomcat fails to complete the Statement benchmark when the Tomcat ``StatementFinalizer`` is used [due to excessive garbage collection times.](https://raw.githubusercontent.com/wiki/brettwooldridge/HikariCP/markdown/Tomcat-Statement-Failure.md)
5 Apache DBCP fails to complete the Statement benchmark [due to excessive garbage collection times](https://raw.githubusercontent.com/wiki/brettwooldridge/HikariCP/markdown/Dbcp2-Statement-Failure.md).
Anaylsis of HikariCP v2.6, in comparison to other pools, in relation to a unique "spike demand" load.
The customer's environment imposed a high cost of new connection acquisition, and a requirement for a dynamically-sized pool, but yet a need for responsiveness to request spikes. Read about the spike demand handling here.
AKA "What you probably didn't know about connection pool sizing". Watch a video from the Oracle Real-world Performance group, and learn about why connection pools do not need to be sized as large as they often are. In fact, oversized connection pools have a clear and demonstrable negative impact on performance; a 50x difference in the case of the Oracle demonstration. Read on to find out..
We'd like to thank the guys over at WIX for the unsolicited and deep write-up about HikariCP on their engineering blog. Take a look if you have time.
Read our interesting "Database down" pool challenge.
Open source software like HikariCP, like any product, competes in the free market. We get it. We understand that product advancements, once public, are often co-opted. And we understand that ideas can arise from the zeitgeist; simultaneously and independently. But the timeline of innovation, particularly in open source projects, is also clear and we want our users to understand the direction of flow of innovation in our space. It could be demoralizing to see the result of hundreds of hours of thought and research co-opted so easily, and perhaps that is inherent in a free marketplace, but we are not demoralized. We are motivated; to widen the gap.
HikariCP comes with sane defaults that perform well in most deployments without additional tweaking. Every property is optional, except for the "essentials" marked below.
📎 HikariCP uses milliseconds for all time values.
🚨 HikariCP relies heavily on accurate high-resolution timers for both performance and reliability. It is imperative that your server is synchronized with a time-source such as an NTP server. Especially if your server is running within a virtual machine. Do not rely on hypervisor settings to "synchronize" the clock of the virtual machine. Configure time-source synchronization inside the virtual machine. If you come asking for support on an issue that turns out to be caused by lack time synchronization, you will be taunted publicly on Twitter.
🔠dataSourceClassName
This is the name of the DataSource
class provided by the JDBC driver. Consult the
documentation for your specific JDBC driver to get this class name, or see the table below.
Note XA data sources are not supported. XA requires a real transaction manager like
bitronix. Note that you do not need this property if you are using
jdbcUrl
for "old-school" DriverManager-based JDBC driver configuration.
Default: none
- or -
🔠jdbcUrl
This property directs HikariCP to use "DriverManager-based" configuration. We feel that DataSource-based
configuration (above) is superior for a variety of reasons (see below), but for many deployments there is
little significant difference. When using this property with "old" drivers, you may also need to set
the driverClassName
property, but try it first without. Note that if this property is used, you may
still use DataSource properties to configure your driver and is in fact recommended over driver parameters
specified in the URL itself.
Default: none
🔠username
This property sets the default authentication username used when obtaining Connections from
the underlying driver. Note that for DataSources this works in a very deterministic fashion by
calling DataSource.getConnection(*username*, password)
on the underlying DataSource. However,
for Driver-based configurations, every driver is different. In the case of Driver-based, HikariCP
will use this username
property to set a user
property in the Properties
passed to the
driver's DriverManager.getConnection(jdbcUrl, props)
call. If this is not what you need,
skip this method entirely and call addDataSourceProperty("username", ...)
, for example.
Default: none
🔠password
This property sets the default authentication password used when obtaining Connections from
the underlying driver. Note that for DataSources this works in a very deterministic fashion by
calling DataSource.getConnection(username, *password*)
on the underlying DataSource. However,
for Driver-based configurations, every driver is different. In the case of Driver-based, HikariCP
will use this password
property to set a password
property in the Properties
passed to the
driver's DriverManager.getConnection(jdbcUrl, props)
call. If this is not what you need,
skip this method entirely and call addDataSourceProperty("pass", ...)
, for example.
Default: none
✅autoCommit
This property controls the default auto-commit behavior of connections returned from the pool.
It is a boolean value.
Default: true
⌚connectionTimeout
This property controls the maximum number of milliseconds that a client (that's you) will wait
for a connection from the pool. If this time is exceeded without a connection becoming
available, a SQLException will be thrown. Lowest acceptable connection timeout is 250 ms.
Default: 30000 (30 seconds)
⌚idleTimeout
This property controls the maximum amount of time that a connection is allowed to sit idle in the
pool. This setting only applies when minimumIdle
is defined to be less than maximumPoolSize
.
Whether a connection is retired as idle or not is subject to a maximum variation of +30
seconds, and average variation of +15 seconds. A connection will never be retired as idle before
this timeout. A value of 0 means that idle connections are never removed from the pool. The mimimum
allowed value is 10000ms (10 seconds).
Default: 600000 (10 minutes)
⌚maxLifetime
This property controls the maximum lifetime of a connection in the pool. An in-use connection will
never be retired, only when it is closed will it then be removed. We strongly recommend setting
this value, and it should be at least 30 seconds less than any database or infrastructure imposed
connection time limit. A value of 0 indicates no maximum lifetime (infinite lifetime), subject of
course to the idleTimeout
setting.
Default: 1800000 (30 minutes)
🔠connectionTestQuery
If your driver supports JDBC4 we strongly recommend not setting this property. This is for
"legacy" databases that do not support the JDBC4 Connection.isValid() API
. This is the query that
will be executed just before a connection is given to you from the pool to validate that the
connection to the database is still alive. Again, try running the pool without this property,
HikariCP will log an error if your driver is not JDBC4 compliant to let you know.
Default: none
🔢minimumIdle
This property controls the minimum number of idle connections that HikariCP tries to maintain
in the pool. If the idle connections dip below this value, HikariCP will make a best effort to
add additional connections quickly and efficiently. However, for maximum performance and
responsiveness to spike demands, we recommend not setting this value and instead allowing
HikariCP to act as a fixed size connection pool.
Default: same as maximumPoolSize
🔢maximumPoolSize
This property controls the maximum size that the pool is allowed to reach, including both
idle and in-use connections. Basically this value will determine the maximum number of
actual connections to the database backend. A reasonable value for this is best determined
by your execution environment. When the pool reaches this size, and no idle connections are
available, calls to getConnection() will block for up to connectionTimeout
milliseconds
before timing out.
Default: 10
📈metricRegistry
This property is only available via programmatic configuration or IoC container. This property
allows you to specify an instance of a Codahale/Dropwizard MetricRegistry
to be used by the
pool to record various metrics. See the Metrics
wiki page for details.
Default: none
📈healthCheckRegistry
This property is only available via programmatic configuration or IoC container. This property
allows you to specify an instance of a Codahale/Dropwizard HealthCheckRegistry
to be used by the
pool to report current health information. See the Health Checks
wiki page for details.
Default: none
🔠poolName
This property represents a user-defined name for the connection pool and appears mainly
in logging and JMX management consoles to identify pools and pool configurations.
Default: auto-generated
⌚initializationFailTimeout
This property controls whether the pool will "fail fast" if the pool cannot be seeded with
an initial connection successfully. Any positive number is taken to be the number of
milliseconds to attempt to acquire an initial connection; the application thread will be
blocked during this period. If a connection cannot be acquired before this timeout occurs,
an exception will be thrown. This timeout is applied after the connectionTimeout
period. If the value is zero (0), HikariCP will attempt to obtain and validate a connection.
If a connection is obtained, but fails validation, an exception will be thrown and the pool
not started. However, if a connection cannot be obtained, the pool will start, but later
efforts to obtain a connection may fail. A value less than zero will bypass any initial
connection attempt, and the pool will start immediately while trying to obtain connections
in the background. Consequently, later efforts to obtain a connection may fail.
Default: 1
❎isolateInternalQueries
This property determines whether HikariCP isolates internal pool queries, such as the
connection alive test, in their own transaction. Since these are typically read-only
queries, it is rarely necessary to encapsulate them in their own transaction. This
property only applies if autoCommit
is disabled.
Default: false
❎allowPoolSuspension
This property controls whether the pool can be suspended and resumed through JMX. This is
useful for certain failover automation scenarios. When the pool is suspended, calls to
getConnection()
will not timeout and will be held until the pool is resumed.
Default: false
❎readOnly
This property controls whether Connections obtained from the pool are in read-only mode by
default. Note some databases do not support the concept of read-only mode, while others provide
query optimizations when the Connection is set to read-only. Whether you need this property
or not will depend largely on your application and database.
Default: false
❎registerMbeans
This property controls whether or not JMX Management Beans ("MBeans") are registered or not.
Default: false
🔠catalog
This property sets the default catalog for databases that support the concept of catalogs.
If this property is not specified, the default catalog defined by the JDBC driver is used.
Default: driver default
🔠connectionInitSql
This property sets a SQL statement that will be executed after every new connection creation
before adding it to the pool. If this SQL is not valid or throws an exception, it will be
treated as a connection failure and the standard retry logic will be followed.
Default: none
🔠driverClassName
HikariCP will attempt to resolve a driver through the DriverManager based solely on the jdbcUrl
,
but for some older drivers the driverClassName
must also be specified. Omit this property unless
you get an obvious error message indicating that the driver was not found.
Default: none
🔠transactionIsolation
This property controls the default transaction isolation level of connections returned from
the pool. If this property is not specified, the default transaction isolation level defined
by the JDBC driver is used. Only use this property if you have specific isolation requirements that are
common for all queries. The value of this property is the constant name from the Connection
class such as TRANSACTION_READ_COMMITTED
, TRANSACTION_REPEATABLE_READ
, etc.
Default: driver default
⌚validationTimeout
This property controls the maximum amount of time that a connection will be tested for aliveness.
This value must be less than the connectionTimeout
. Lowest acceptable validation timeout is 250 ms.
Default: 5000
⌚leakDetectionThreshold
This property controls the amount of time that a connection can be out of the pool before a
message is logged indicating a possible connection leak. A value of 0 means leak detection
is disabled. Lowest acceptable value for enabling leak detection is 2000 (2 seconds).
Default: 0
➡dataSource
This property is only available via programmatic configuration or IoC container. This property
allows you to directly set the instance of the DataSource
to be wrapped by the pool, rather than
having HikariCP construct it via reflection. This can be useful in some dependency injection
frameworks. When this property is specified, the dataSourceClassName
property and all
DataSource-specific properties will be ignored.
Default: none
➡threadFactory
This property is only available via programmatic configuration or IoC container. This property
allows you to set the instance of the java.util.concurrent.ThreadFactory
that will be used
for creating all threads used by the pool. It is needed in some restricted execution environments
where threads can only be created through a ThreadFactory
provided by the application container.
Default: none
➡scheduledExecutor
This property is only available via programmatic configuration or IoC container. This property
allows you to set the instance of the java.util.concurrent.ScheduledExecutorService
that will
be used for various internally scheduled tasks. If supplying HikariCP with a ScheduledThreadPoolExecutor
instance, it is recommended that setRemoveOnCancelPolicy(true)
is used.
Default: none
HikariCP has plenty of "knobs" to turn as you can see above, but comparatively less than some other pools. This is a design philosophy. The HikariCP design aesthetic is Minimalism. In keeping with the simple is better or less is more design philosophy, some configuration axis are intentionally left out.
Many connection pools, including Apache DBCP, Vibur, c3p0 and others offer PreparedStatement
caching.
HikariCP does not. Why?
At the connection pool layer PreparedStatements
can only be cached per connection. If your application
has 250 commonly executed queries and a pool of 20 connections you are asking your database to hold on to
5000 query execution plans -- and similarly the pool must cache this many PreparedStatements
and their
related graph of objects.
Most major database JDBC drivers already have a Statement cache that can be configured, including PostgreSQL,
Oracle, Derby, MySQL, DB2, and many others. JDBC drivers are in a unique position to exploit database specific
features, and nearly all of the caching implementations are capable of sharing execution plans across connections.
This means that instead of 5000 statements in memory and associated execution plans, your 250 commonly executed
queries result in exactly 250 execution plans in the database. Clever implementations do not even retain
PreparedStatement
objects in memory at the driver-level but instead merely attach new instances to existing plan IDs.
Using a statement cache at the pooling layer is an anti-pattern, and will negatively impact your application performance compared to driver-provided caches.
Like Statement caching, most major database vendors support statement logging through properties of their own driver. This includes Oracle, MySQL, Derby, MSSQL, and others. Some even support slow query logging. For those few databases that do not support it, log4jdbc or jdbcdslog-exp are good options.
You can use the HikariConfig
class like so:
HikariConfig config = new HikariConfig();
config.setJdbcUrl("jdbc:mysql://localhost:3306/simpsons");
config.setUsername("bart");
config.setPassword("51mp50n");
config.addDataSourceProperty("cachePrepStmts", "true");
config.addDataSourceProperty("prepStmtCacheSize", "250");
config.addDataSourceProperty("prepStmtCacheSqlLimit", "2048");
HikariDataSource ds = new HikariDataSource(config);
or directly instantiate a HikariDataSource
like so:
HikariDataSource ds = new HikariDataSource();
ds.setJdbcUrl("jdbc:mysql://localhost:3306/simpsons");
ds.setUsername("bart");
ds.setPassword("51mp50n");
...
or property file based:
// Examines both filesystem and classpath for .properties file
HikariConfig config = new HikariConfig("/some/path/hikari.properties");
HikariDataSource ds = new HikariDataSource(config);
Example property file:
dataSourceClassName=org.postgresql.ds.PGSimpleDataSource
dataSource.user=test
dataSource.password=test
dataSource.databaseName=mydb
dataSource.portNumber=5432
dataSource.serverName=localhost
or java.util.Properties
based:
Properties props = new Properties();
props.setProperty("dataSourceClassName", "org.postgresql.ds.PGSimpleDataSource");
props.setProperty("dataSource.user", "test");
props.setProperty("dataSource.password", "test");
props.setProperty("dataSource.databaseName", "mydb");
props.put("dataSource.logWriter", new PrintWriter(System.out));
HikariConfig config = new HikariConfig(props);
HikariDataSource ds = new HikariDataSource(config);
There is also a System property available, hikaricp.configurationFile
, that can be used to specify the
location of a properties file. If you intend to use this option, construct a HikariConfig
or HikariDataSource
instance using the default constructor and the properties file will be loaded.
We recommended using dataSourceClassName
instead of jdbcUrl
, but either is acceptable. We'll say that again, either is acceptable.
⚠ Note: Spring Boot auto-configuration users, you need to use jdbcUrl
-based configuration.
⚠ MysqlDataSource is known to be broken, use jdbcUrl
configuration instead.
Here is a list of JDBC DataSource classes for popular databases:
Database | Driver | DataSource class |
---|---|---|
Apache Derby | Derby | org.apache.derby.jdbc.ClientDataSource |
Firebird | Jaybird | org.firebirdsql.pool.FBSimpleDataSource |
H2 | H2 | org.h2.jdbcx.JdbcDataSource |
HSQLDB | HSQLDB | org.hsqldb.jdbc.JDBCDataSource |
IBM DB2 | IBM JCC | com.ibm.db2.jcc.DB2SimpleDataSource |
IBM Informix | IBM Informix | com.informix.jdbcx.IfxDataSource |
MS SQL Server | Microsoft | com.microsoft.sqlserver.jdbc.SQLServerDataSource |
MySQL | Connector/J | |
MySQL/MariaDB | MariaDB | org.mariadb.jdbc.MySQLDataSource |
Oracle | Oracle | oracle.jdbc.pool.OracleDataSource |
OrientDB | OrientDB | com.orientechnologies.orient.jdbc.OrientDataSource |
PostgreSQL | pgjdbc-ng | com.impossibl.postgres.jdbc.PGDataSource |
PostgreSQL | PostgreSQL | org.postgresql.ds.PGSimpleDataSource |
SAP MaxDB | SAP | com.sap.dbtech.jdbc.DriverSapDB |
SQLite | xerial | org.sqlite.SQLiteDataSource |
SyBase | jConnect | com.sybase.jdbc4.jdbc.SybDataSource |
Note Play 2.4 now uses HikariCP by default. A new plugin has come up for the the Play framework; play-hikaricp. If you're using the excellent Play framework, your application deserves HikariCP. Thanks Edulify Team!
A new Clojure wrapper has been created by tomekw and can be found here.
A new JRuby wrapper has been created by tomekw and can be found here.
Google discussion group HikariCP here, growing FAQ.
Don't forget the Wiki for additional information such as:
⇒ Java 8+ (Java 6/7 artifacts are in maintenance mode)
⇒ slf4j library
YourKit supports open source projects with its full-featured Java Profiler. Click the YourKit logo below to learn more.
Please perform changes and submit pull requests from the dev
branch instead of master
. Please set your editor to use spaces instead of tabs, and adhere to the apparent style of the code you are editing. The dev
branch is always more "current" than the master
if you are looking to live life on the edge.