Abstract: Working set ratio estimations of data items in a sliding time window are determined to dynamically allocate storage for the data items. A working set ratio may be determined by accessing a fixed-size array that stores respective timestamps of last accesses of data items to determine which data items are useful to determine an estimate of a working set for the application within a range of time. The working set ratio is then determined from an estimated working set and an amount of computing resources allocated to the application by the estimated working set. The amount of the computing resources allocated to the application may then be automatically scaled according to the determine working set ratio.

Abstract: Intelligent query routing may be performed across shards of a scalable database table. A router of a database system may receive an access request directed to one or more database tables. The router may evaluate the access request with respect to metadata obtained for the database tables to determine an assignment distribution of computing resources of the database system to data that can satisfy the access request. The router can select planning locations to perform the access request based on the assignment distribution of the computing resources. The router can cause the access request to be performed according to planning at the selected planning locations.

Abstract: Lease-based consistency may be implemented for databases to handle failovers. A database node may obtain a consistency lease that describes a point in time determined from a time-to-live amount added to a consistent point in time for database data. While the consistency lease is valid, Multi-version Concurrency Control (MVCC) snapshots assigned by the database node can be used to handle requests to access the database data. Once expired, the database node may have to renew the consistency lease in order to continue to handle write and read requests.

Abstract: Techniques for aggregating and emitting database activity record batches are described. Database activity records can be written to a shared memory queue and emitted to a destination using a remote procedure call (RPC). Individual database connection server processes can write client activity records to the queue. An activity monitor plugin in the database engine can monitor the audit records and aggregate the audit records into batches. Batches of audit records can be sent via RPC to their final or intermediate destination. Each instance host in a database service can include a client backend process configured to define how to submit audit records to shared memory. The activity monitor plugin can batch audit records in to messages and submit those messages via RPC to a security host manager and relaying response back to each relevant client backend.