Endpoints and connectivity for Timestream for InfluxDB 3

Amazon Timestream for InfluxDB 3 provides multiple endpoints to connect to your cluster, allowing for flexible access patterns based on your application needs.

Endpoint types

Cluster endpoint

The cluster endpoint (or primary endpoint) provides access to the writer nodes in your cluster. This endpoint:

• Supports both read and write operations.

• Automatically routes traffic to available writer nodes.

• Is the only endpoint that can perform write operations.

• Should be used for administrative operations and data ingestion.

Example format: clusterid-wmyjrrjko.timestream-influxdb-alpha.us-west-2.on.aws

Reader endpoint

The reader endpoint connects to the reader nodes in your cluster. This endpoint:

• Supports read-only operations (queries).

• Automatically distributes read traffic across all available reader nodes.

• Helps offload query workload from writer nodes.

• Is ideal for reporting and dashboard applications.

Example format: clusterid-wmyjrrjko-ro.timestream-influxdb-alpha.us-west-2.on.aws

Node-specific endpoints

In addition to the cluster endpoints, you can connect directly to specific nodes within your cluster:

• Provides direct access to individual nodes for diagnostics or specific workloads.

• Allows for fine-grained control over connection routing.

• Useful for troubleshooting or when you need to isolate specific operations.

• Important: During any issue or failover scenario, node-specific endpoints will be temporarily unavailable while the node recovers or is replaced.

Example format: nodeid-wmyjrrjko.timestream-influxdb-alpha.us-west-2.on.aws

Traffic distribution

When using cluster-level endpoints (writer/reader and reader endpoints), the traffic distribution system automatically:

• Distributes traffic to active nodes based on:

  • Node health and availability.

  • Current workload on each node.

  • Node role (writer/reader).

• Makes specific node unavailability transparent to clients by automatically routing traffic away from unhealthy or unavailable nodes.

• Provides seamless failover without requiring client-side configuration changes.

• Ensures continuous service availability even during maintenance or unexpected node failures.

This automatic traffic distribution provides:

• High availability: Client connections remain uninterrupted even when individual nodes become unavailable.

• Load balancing: Traffic is distributed across available nodes for optimal performance.

• Fault tolerance: Automatic failover ensures minimal impact during node failures.

• Simplified connection management: Applications don't need to handle node-level failures.

Connection best practices

For optimal performance and reliability:

• Use cluster-level endpoints for production workloads: The writer/reader and reader endpoints provide automatic failover and transparent handling of node unavailability.

• Use the writer/reader endpoint for write operations and administrative tasks.

• Use the reader endpoint for read-heavy applications like dashboards and reports.

• Avoid node-specific endpoints for critical applications: These endpoints will be unavailable during failover scenarios and don't provide automatic traffic redistribution.

• Reserve node-specific endpoints for diagnostics, troubleshooting, or when you need to isolate specific operations.

• Implement connection pooling in your applications to efficiently manage connections.

• Configure appropriate timeouts and retry logic in your client applications.

By leveraging the cluster-level endpoints and their automatic traffic distribution capabilities, you can ensure your applications maintain continuous connectivity and optimal performance even during node-level failures or maintenance events.