About configuration in TimescaleDB

By default, TimescaleDB uses the default PostgreSQL server configuration settings. However, in some cases, these settings are not appropriate, especially if you have larger servers that use more hardware resources such as CPU, memory, and storage. This section explains some of the settings you are most likely to need to adjust.

Some of these settings are PostgreSQL settings, and some are TimescaleDB specific settings. For most changes, you can use the tuning tool to adjust your configuration. For more advanced configuration settings, or to change settings that aren't included in the timescaledb-tune tool, you can manually adjust the postgresql.conf configuration file.

Settings:

• shared_buffers
• effective_cache_size
• work_mem
• maintenance_work_mem
• max_connections

You can adjust each of these to match the machine's available memory. To make it easier, you can use the PgTune site to work out what settings to use: enter your machine details, and select the data warehouse DB type to see the suggested parameters.

Tip

You can adjust these settings with timescaledb-tune.

Settings:

• timescaledb.max_background_workers
• max_parallel_workers
• max_worker_processes

PostgreSQL uses worker pools to provide workers for live queries and background jobs. If you do not configure these settings, your queries and background jobs could run more slowly.

TimescaleDB background workers are configured with timescaledb.max_background_workers. Each database needs a background worker allocated to schedule jobs. Additional workers run background jobs as required. This setting should be the sum of the total number of databases and the total number of concurrent background workers you want running at any one time. By default, timescaledb-tune sets timescaledb.max_background_workers to 16. You can change this setting directly, use the --max-bg-workers flag, or adjust the TS_TUNE_MAX_BG_WORKERS Docker environment variable.

TimescaleDB parallel workers are configured with max_parallel_workers. For larger queries, PostgreSQL automatically uses parallel workers if they are available. Increasing this setting can improve query performance for large queries that trigger the use of parallel workers. By default, this setting corresponds to the number of CPUs available. You can change this parameter directly, by adjusting the --cpus flag, or by using the TS_TUNE_NUM_CPUS Docker environment variable.

The max_worker_processes setting defines the total pool of workers available to both background and parallel workers, as well a small number of built-in PostgreSQL workers. It should be at least the sum of timescaledb.max_background_workers and max_parallel_workers.

Tip

You can adjust these settings with timescaledb-tune.

Settings:

• synchronous_commit

By default, disk writes are performed synchronously, so each transaction must be completed and a success message sent, before the next transaction can begin. You can change this to asynchronous to increase write throughput by setting synchronous_commit = 'off'. Note that disabling synchronous commits could result in some committed transactions being lost. To help reduce the risk, do not also change fsync setting. For more information about asynchronous commits and disk write speed, see the PostgreSQL documentation.

Tip

You can adjust these settings in the postgresql.conf configuration file.

Settings:

• max_locks_per_transaction

TimescaleDB relies on table partitioning to scale time-series workloads. A hypertable needs to acquire locks on many chunks during queries, which can exhaust the default limits for the number of allowed locks held. In some cases, you might see a warning like this:

psql: FATAL:  out of shared memory
HINT:  You might need to increase max_locks_per_transaction.

Copy

To avoid this issue, you can increase the max_locks_per_transaction setting from the default value, which is usually 64. This parameter limits the average number of object locks used by each transaction; individual transactions can lock more objects as long as the locks of all transactions fit in the lock table.

For most workloads, choose a number equal to double the maximum number of chunks you expect to have in a hypertable divided by max_connections. This takes into account that the number of locks used by a hypertable query is roughly equal to the number of chunks in the hypertable if you need to access all chunks in a query, or double that number if the query uses an index. You can see how many chunks you currently have using the timescaledb_information.hypertables view. Changing this parameter requires a database restart, so make sure you pick a larger number to allow for some growth. For more information about lock management, see the PostgreSQL documentation.

Tip

You can adjust these settings in the postgresql.conf configuration file.