Migrating to Amazon Timestream for InfluxDB 3

Amazon Timestream for InfluxDB 3 Enterprise is the recommended migration target for most LiveAnalytics workloads. It provides standard SQL and InfluxQL query support, unlimited cardinality, multi-node scaling, and a built-in Python processing engine.

Key differences from LiveAnalytics

Feature	LiveAnalytics	InfluxDB 3 Enterprise
Architecture	Serverless, fully managed	Instance-based clusters (db.influx.* family)
Storage	Automatic memory + magnetic tiers	Amazon S3 object storage (Apache Parquet)
Query language	Custom Timestream SQL dialect	Standard SQL + InfluxQL
Write protocol	AWS SDK (WriteRecords API)	Line Protocol over HTTPS (port 443)
Scaling	Automatic	Manual: choose instance size + node count
Cardinality	Automatic partitioning	Unlimited (no series key index)
Processing	Scheduled queries	Built-in Python processing engine (plugins, triggers)
Auth model	IAM for data-plane	Token-based authentication (Admin Token at provisioning)
Pricing	Per-write, per-query, per-GB stored	Per-instance-hour + included Amazon S3 I/O + InfluxData license (Marketplace)
Availability SLA	99.99%	99.9% (multi-node, multi-AZ)
Regions	10 regions (frozen)	19+ regions (expanding)

Instance types available

Instance	vCPUs	Memory
db.influx.large	2	8 GB
db.influx.xlarge	4	16 GB
db.influx.2xlarge	8	32 GB
db.influx.4xlarge	16	64 GB
db.influx.8xlarge	32	128 GB
db.influx.12xlarge	48	192 GB

Editions

InfluxDB 3 Core

Single-node, open-source engine, no license fee. Good for development, testing, and smaller workloads.

InfluxDB 3 Enterprise

Multi-node clusters (up to 15 nodes), InfluxData Marketplace license required. Recommended for production workloads at scale. Supports configurable node modes (INGEST, QUERY, COMPACT, PROCESS, or ALL).

Feature mapping

LiveAnalytics feature	InfluxDB 3 equivalent
Scheduled Queries (downsampling)	Processing Engine scheduled triggers (Python plugins)
Memory Store (recent data)	WAL buffer + Last Value Cache
Magnetic Store (historical data)	Amazon S3 Parquet files (compacted)
Multi-measure records	Multiple fields per line protocol point
Dimension-based partitioning	Tag-based automatic partitioning
UNLOAD to Amazon S3	Export API (table-level Parquet export)
Cross-region replication	Not available (use multi-AZ within region)
AWS Backup integration	Automated backups (snapshot-based)
VPC Endpoints	Private clusters deployed in customer VPC
AWS KMS encryption	Encryption at rest (AWS managed keys; CMK support available)

Steps to migrate from LiveAnalytics to InfluxDB 3

1. Assess your workload: Identify write throughput (records/sec), query patterns (point queries vs. aggregations), cardinality, and data retention needs. If using Scheduled Queries in LiveAnalytics, plan to replace with InfluxDB 3 Processing Engine plugins.

2. Provision your InfluxDB 3 cluster: Choose Core (single-node) or Enterprise (multi-node) based on throughput needs. Create via AWS Console, AWS CLI, or CloudFormation. Select instance size based on write throughput and query concurrency.

3. Update write path: LiveAnalytics uses the WriteRecords API (AWS SDK). InfluxDB 3 uses Line Protocol over HTTPS. Update Telegraf outputs, application code, or IoT Rules to write via Line Protocol to the cluster endpoint. Batch writes with 5,000–10,000 points per batch, sorted by tag keys alphabetically.

4. Update query path: LiveAnalytics custom SQL maps to InfluxDB 3 standard SQL with key differences: use time column directly (no measure_value::double pattern), use WHERE time > now() - INTERVAL '1 hour' instead of ago(). Use Flight SQL (port 443) for programmatic access or InfluxQL for Grafana dashboards.

5. Migrate historical data: Export from LiveAnalytics using UNLOAD to Amazon S3 (Parquet or CSV). Import into InfluxDB 3 using Line Protocol writes. For the export step, follow the instructions in Exporting Timestream data to Amazon S3. Bulk import from Parquet files is on the roadmap for a future release.

6. Update monitoring: Replace LiveAnalytics CloudWatch metrics with InfluxDB 3 CloudWatch metrics (CPUUtilization, MemoryUtilization, S3ObjectStorageBytes). Additional metrics are available via the /metrics Prometheus endpoint on the cluster.

7. Update networking and auth: LiveAnalytics uses IAM for data-plane auth. InfluxDB 3 uses token-based auth (Admin Token created at cluster provisioning). IAM-based data-plane auth for InfluxDB 3 is not yet available. The cluster must be deployed in a VPC with appropriate security groups.

Additional resources

• Getting started with InfluxDB 3

• InfluxDB 3 DB Clusters

• InfluxDB 3 Processing Engine

• Line Protocol Reference

• SQL Reference for InfluxDB 3