RDS Proxy concepts and terminology
You can simplify connection management for your Amazon RDS DB instances by using RDS Proxy.
RDS Proxy handles the network traffic between the client application and the database. It does so in an active way first by understanding the database protocol. It then adjusts its behavior based on the SQL operations from your application and the result sets from the database.
RDS Proxy reduces the memory and CPU overhead for connection management on your database. The database needs less memory and CPU resources when applications open many simultaneous connections. It also doesn't require logic in your applications to close and reopen connections that stay idle for a long time. Similarly, it requires less application logic to reestablish connections in case of a database problem.
The infrastructure for RDS Proxy is highly available and deployed over multiple Availability Zones (AZs). The computation, memory, and storage for RDS Proxy are independent of your RDS DB instance. This separation helps lower overhead on your database servers, so that they can devote their resources to serving database workloads. The RDS Proxy compute resources are serverless, automatically scaling based on your database workload.
Overview of RDS Proxy concepts
RDS Proxy handles the infrastructure to perform connection pooling and the other features described in the sections that follow. You see the proxies represented in the RDS console on the Proxies page.
Each proxy handles connections to a single RDS DB instance. The proxy automatically determines the current writer instance for RDS Multi-AZ DB instance or cluster.
The connections that a proxy keeps open and available for your database applications to use make up the connection pool.
By default, RDS Proxy can reuse a connection after each transaction in your session. This transaction-level reuse is called multiplexing. When RDS Proxy temporarily removes a connection from the connection pool to reuse it, that operation is called borrowing the connection. When it's safe to do so, RDS Proxy returns that connection to the connection pool.
In some cases, RDS Proxy can't be sure that it's safe to reuse a database connection outside of the current session. In these cases, it keeps the session on the same connection until the session ends. This fallback behavior is called pinning.
A proxy has a default endpoint. You connect to this endpoint when you work with an Amazon RDS DB instance. You do so instead of connecting to the read/write endpoint that connects directly to the instance . For RDS DB clusters, you can also create additional read/write and read-only endpoints. For more information, see Overview of proxy endpoints.
For example, you can still connect to the cluster endpoint for read/write connections without connection pooling. You can still connect to the reader endpoint for load-balanced read-only connections. You can still connect to the instance endpoints for diagnosis and troubleshooting of specific DB instances within a cluster. If you use other Amazon services such as Amazon Lambda to connect to RDS databases, change their connection settings to use the proxy endpoint. For example, you specify the proxy endpoint to allow Lambda functions to access your database while taking advantage of RDS Proxy functionality.
Each proxy contains a target group. This target group embodies the RDS DB instance that the proxy can connect to. The RDS DB instance associated with a proxy are called the targets of that proxy. For convenience, when you create a proxy through the console, RDS Proxy also creates the corresponding target group and registers the associated targets automatically.
An engine family is a related set of database engines that use the same DB protocol. You choose the engine family for each proxy that you create.
Connection pooling
Each proxy performs connection pooling for the writer instance of its associated RDS database . Connection pooling is an optimization that reduces the overhead associated with opening and closing connections and with keeping many connections open simultaneously. This overhead includes memory needed to handle each new connection. It also involves CPU overhead to close each connection and open a new one. Examples include Transport Layer Security/Secure Sockets Layer (TLS/SSL) handshaking, authentication, negotiating capabilities, and so on. Connection pooling simplifies your application logic. You don't need to write application code to minimize the number of simultaneous open connections.
Each proxy also performs connection multiplexing, also known as connection reuse. With multiplexing, RDS Proxy performs all the operations for a transaction using one underlying database connection. RDS then can use a different connection for the next transaction. You can open many simultaneous connections to the proxy, and the proxy keeps a smaller number of connections open to the DB instance or cluster. Doing so further minimizes the memory overhead for connections on the database server. This technique also reduces the chance of "too many connections" errors.
RDS Proxy security
RDS Proxy uses the existing RDS security mechanisms such as TLS/SSL and Amazon Identity and Access Management (IAM). For general information about those security features, see Security in Amazon RDS. Also, make sure to familiarize yourself with how RDS work with authentication, authorization, and other areas of security.
RDS Proxy can act as an additional layer of security between client applications and the underlying database. For example, you can connect to the proxy using TLS 1.3, even if the underlying DB instance supports an older version of TLS. You can connect to the proxy using an IAM role. This is so even if the proxy connects to the database using the native user and password authentication method. By using this technique, you can enforce strong authentication requirements for database applications without a costly migration effort for the DB instances themselves.
You store the database credentials used by RDS Proxy in Amazon Secrets Manager. Each database user for the RDS DB instance accessed by a proxy must have a corresponding secret in Secrets Manager. You can also set up IAM authentication for users of RDS Proxy. By doing so, you can enforce IAM authentication for database access even if the databases use native password authentication. We recommend using these security features instead of embedding database credentials in your application code.
Using TLS/SSL with RDS Proxy
You can connect to RDS Proxy using the TLS/SSL protocol.
Note
RDS Proxy uses certificates from the Amazon Certificate Manager (ACM). If you are using RDS Proxy, you don't need to download Amazon RDS certificates or update applications that use RDS Proxy connections.
To enforce TLS for all connections between the proxy and your database, you can specify a setting Require Transport Layer Security when you create or modify a proxy in the Amazon Web Services Management Console.
RDS Proxy can also ensure that your session uses TLS/SSL between your client and the RDS Proxy endpoint. To have RDS Proxy do so, specify the requirement on the client side. SSL session variables are not set for SSL connections to a database using RDS Proxy.
-
For RDS for MySQL, specify the requirement on the client side with the
--ssl-mode
parameter when you run themysql
command. -
For Amazon RDS PostgreSQL, specify
sslmode=require
as part of theconninfo
string when you run thepsql
command.
RDS Proxy supports TLS protocol version 1.0, 1.1, 1.2, and 1.3. You can connect to the proxy using a higher version of TLS than you use in the underlying database.
By default, client programs establish an encrypted connection with RDS Proxy, with
further control available through the --ssl-mode
option. From the client
side, RDS Proxy supports all SSL modes.
For the client, the SSL modes are the following:
- PREFERRED
-
SSL is the first choice, but it isn't required.
- DISABLED
-
No SSL is allowed.
- REQUIRED
-
Enforce SSL.
- VERIFY_CA
-
Enforce SSL and verify the certificate authority (CA).
- VERIFY_IDENTITY
-
Enforce SSL and verify the CA and CA hostname.
When using a client with --ssl-mode
VERIFY_CA
or VERIFY_IDENTITY
, specify the --ssl-ca
option pointing to a CA in .pem
format. For the .pem
file to
use, download all root CA PEMs from Amazon Trust Services .pem
file.
RDS Proxy uses wildcard certificates, which apply to both a domain and its subdomains. If you use the
mysql
client to connect with SSL mode VERIFY_IDENTITY
, currently you must use
the MySQL 8.0-compatible mysql
command.
Failover
Failover is a high-availability feature that replaces a database instance with another one when the original instance becomes unavailable. A failover might happen because of a problem with a database instance. It might also be part of normal maintenance procedures, such as during a database upgrade. Failover applies to RDS DB instances in a Multi-AZ configuration.
Connecting through a proxy makes your applications more resilient to database failovers. When the original DB instance becomes unavailable, RDS Proxy connects to the standby database without dropping idle application connections. This helps speed up and simplify the failover process. This is less disruptive to your application than a typical reboot or database problem.
Without RDS Proxy, a failover involves a brief outage. During the outage, you can't perform write operations on the database in failover. Any existing database connections are disrupted, and your application must reopen them. The database becomes available for new connections and write operations when a read-only DB instance is promoted in place of one that's unavailable.
During DB failovers, RDS Proxy continues to accept connections at the same IP address and automatically directs connections to the new primary DB instance. Clients connecting through RDS Proxy are not susceptible to the following:
-
Domain Name System (DNS) propagation delays on failover.
-
Local DNS caching.
-
Connection timeouts.
-
Uncertainty about which DB instance is the current writer.
-
Waiting for a query response from a former writer that became unavailable without closing connections.
For applications that maintain their own connection pool, going through RDS Proxy means that most connections stay alive during failovers or other disruptions. Only connections that are in the middle of a transaction or SQL statement are canceled. RDS Proxy immediately accepts new connections. When the database writer is unavailable, RDS Proxy queues up incoming requests.
For applications that don't maintain their own connection pools, RDS Proxy offers faster connection rates and more open connections. It offloads the expensive overhead of frequent reconnects from the database. It does so by reusing database connections maintained in the RDS Proxy connection pool. This approach is particularly important for TLS connections, where setup costs are significant.
Transactions
All the statements within a single transaction always use the same underlying database connection. The connection becomes available for use by a different session when the transaction ends. Using the transaction as the unit of granularity has the following consequences:
-
Connection reuse can happen after each individual statement when the RDS for MySQL
autocommit
setting is turned on. -
Conversely, when the
autocommit
setting is turned off, the first statement you issue in a session begins a new transaction. For example, suppose that you enter a sequence ofSELECT
,INSERT
,UPDATE
, and other data manipulation language (DML) statements. In this case, connection reuse doesn't happen until you issue aCOMMIT
,ROLLBACK
, or otherwise end the transaction. -
Entering a data definition language (DDL) statement causes the transaction to end after that statement completes.
RDS Proxy detects when a transaction ends through the network protocol used by the database client
application. Transaction detection doesn't rely on keywords such as COMMIT
or
ROLLBACK
appearing in the text of the SQL statement.
In some cases, RDS Proxy might detect a database request that makes it impractical to move your session to a different connection. In these cases, it turns off multiplexing for that connection the remainder of your session. The same rule applies if RDS Proxy can't be certain that multiplexing is practical for the session. This operation is called pinning. For ways to detect and minimize pinning, see Avoiding pinning an RDS Proxy.