Prompt
I’m unhappy with AWS Timestream pricing for continuous ingestion
Latest observation
Oct 25, 2026
OpenAI APIWeb search: off
If you’re unhappy with AWS Timestream pricing for continuous ingestion, the main lever is usually to reduce write volume and retained “hot” data, or consider a cheaper time-series path if your workload is steady and high-throughput.
A few practical options:
1) Tune Timestream usage
- Reduce ingestion frequency: batch points instead of writing every event.
- Aggregate before writing: write 1-min or 5-min rollups instead of raw events where possible.
- Shorten memory retention: keep only the hottest data in memory, move to magnetic storage sooner.
- Filter noisy metrics: drop low-value dimensions/duplicate measurements.
- Use multi-measure records: combine several metrics into one record to reduce write count.
2) Re-evaluate the architecture
If you need continuous high-volume ingestion, Timestream may not be the cheapest fit. Alternatives depend on query pattern:
- Amazon S3 + Athena / Iceberg / Glue
Cheapest for long-term storage and batch/analytical queries, but not ideal for low-latency time-series reads. - Amazon OpenSearch
Good for dashboards and near-real-time queries, but can get expensive at scale. - PostgreSQL/TimescaleDB
Strong option for time-series workloads if you want more predictable pricing and SQL. - ClickHouse
Often very cost-effective for high-ingest analytics and time-series-style workloads. - InfluxDB / VictoriaMetrics
Purpose-built time-series databases; pricing can be more predictable depending on deployment.
3) Hybrid approach
A common pattern:
- Raw events → S3
- Aggregates / recent operational data → Timestream or another fast store
- Historical analytics → Athena/Glue/Iceberg or warehouse
This can dramatically cut Timestream ingestion costs.
4) If staying on AWS
You could also look at:
- Kinesis Firehose to S3 for cheaper ingestion/storage
- Lambda / Flink / ECS to pre-aggregate before storing
- Athena over partitioned files for historical analysis
If you want, I can help you with one of these:
- Estimate whether Timestream is overpriced for your workload
- Suggest a cheaper AWS-native architecture
- Recommend a non-AWS replacement based on your ingest rate and query needs
If you share rough numbers like events/sec, record size, retention, and query latency needs, I can give a much more concrete recommendation.