For similar capabilities to Amazon Timestream for LiveAnalytics, consider Amazon Timestream for InfluxDB. It offers simplified data ingestion and single-digit millisecond query response times for real-time analytics. Learn more here.
Supported Instance Types and Specifications
The following table lists all supported instance types for Amazon Timestream, along with their hardware specifications. These specifications are critical for determining the optimal parameter group values for your workload.
| Instance Type | vCPUs | Memory (GiB) | Memory (Bytes) | Use Case Profile |
|---|---|---|---|---|
| db.influx.medium | 1 | 8 | 8,589,934,592 | Dev/test, low-volume workloads |
| db.influx.large | 2 | 16 | 17,179,869,184 | Light production, small datasets |
| db.influx.xlarge | 4 | 32 | 34,359,738,368 | Small production workloads |
| db.influx.2xlarge | 8 | 64 | 68,719,476,736 | Medium production workloads |
| db.influx.4xlarge | 16 | 128 | 137,438,953,472 | Large production workloads |
| db.influx.8xlarge | 32 | 256 | 274,877,906,944 | High-throughput production |
| db.influx.12xlarge | 48 | 384 | 412,316,860,416 | Heavy analytics workloads |
| db.influx.16xlarge | 64 | 512 | 549,755,813,888 | Maximum r7g capacity |
| db.influx.24xlarge | 96 | 768 | 824,633,720,832 | Extreme-scale workloads |
The following table provides recommended parameter values for each instance size. Parameters not listed use the same default across all sizes. For detailed explanations, see Detailed Parameter Reference.
| Parameter | Medium | Large | XLarge | 2XLarge | 4XLarge | 8XLarge | 12XLarge | 16XLarge | 24XLarge |
|---|---|---|---|---|---|---|---|---|---|
| num-datafusion-threads | 1 | 2 | 4 | 8 | 16 | 32 | 48 | 64 | 96 |
| num-io-threads | 1 | 2 | 4 | 8 | 16 | 32 | 48 | 64 | 96 |
| datafusion-max-parquet-fanout | 250–500 | 500 | 500–1K | 1K | 1K–2K | 2K–5K | 5K–10K | 5K–10K | 5K–10K |
| query-file-limit | 100–200 | 200–300 | 300–432 | 432 | 432–600 | 600–800 | 800–1024 | 800–1024 | 1024 |
| query-log-size | 500 | 1K | 1K | 1K–2K | 2K–5K | 2K–5K | 2K–5K | 2K–5K | 2K–5K |
| exec-mem-pool-bytes | 20% | 20% | 20% | 20% | 20% | 20% | 20% | 20% | 20% |
| force-snapshot-mem-threshold | 30% | 40% | 50% | 50% | 50% | 50% | 50% | 50% | 50% |
| parquet-mem-cache-size | 15% | 20% | 20% | 20% | 25% | 25% | 25% | 25% | 25% |
| wal-max-write-buffer-size | 50K–100K | 100K | 100K–200K | 200K–300K | 300K–500K | 500K–700K | 500K–800K | 700K–1M | 800K–1M |
| wal-snapshot-size | 300–600 | 600 | 600–1K | 1K–2K | 2K–5K | 2K–5K | 2K–5K | 2K–5K | 2K–5K |
| compaction-max-num-files-per-plan (Ent.) | 500 | 500 | 500 | 500–1K | 1K–2K | 2K–5K | 5K | 5K–10K | 5K–10K |
| table-index-cache-concurrency-limit | 8–16 | 16–20 | 20 | 20–24 | 24–32 | 32–48 | 48–64 | 48–64 | 64–100 |
| max-http-request-size | 5 MiB | 10 MiB | 10 MiB | 10–16 MiB | 10–16 MiB | 10–16 MiB | 10–16 MiB | 10–16 MiB | 10–16 MiB |
Note: Parameters not listed in this table (such as datafusion-use-cached-parquet-loader, disable-parquet-mem-cache, parquet-mem-cache-prune-interval, compaction-multipliers, compaction-row-limit, cache eviction intervals, and others) use the same recommended default value across all instance sizes. See Detailed Parameter Reference for their defaults.
Note: Parameters not listed in the table above use the same recommended default value across all instance sizes. See Detailed Parameter Reference for their defaults.
