Compare power choices before counting replicas

A two-node service and a larger single node solve different resource problems. The Lightsail price table, checked on 10 November 2026, lists container prices per node. These are ordinary USD monthly prices, not promotional credits or a whole-application invoice.

Power Published vCPU per node RAM per node USD/month per node
Micro 0.25 shared 1 GB 10
Small 0.5 shared 1 GB 15
Medium 1 2 GB 40

Small raises the CPU allocation above Micro while keeping the same listed memory. If one application process needs more than the available 1 GB-node budget, two Micro nodes do not turn into a single 2 GB node. Medium supplies a different per-node memory size. None of these allocations is an independently tested throughput result.

Multiply the node price by scale

The official FAQ says the maximum monthly service price is the power’s base price multiplied by scale. Its own example is Micro at scale two: USD10 × 2 = USD20/month. Usage is billed hourly up to that monthly maximum.

Compare that USD20 base with the deployment you actually need. Two replicas require the application to distribute requests safely and handle its state deliberately; a memory-heavy single process needs a different power choice. Add external data services, applicable transfer charges and tax separately rather than calling the node subtotal the final bill. This job has not verified any account’s trial or credit eligibility.

Keep the transfer allowance at service level

The FAQ gives each container service a 500 GB monthly allowance. Both inbound and outbound data consume it, with specified excess outbound paths subject to fees. Scaling to two nodes is therefore not evidence of a 1,000 GB allowance.

For a container that imports data then returns results, budget both paths against the service allowance. Use the selected region’s published outbound tariff for the actual billable path; no single global overage total is established here. This service-level rule is distinct from the VM bundle allowances shown in the catalog and should not be replaced with them.

Check the delivery pipeline before choosing managed containers

The FAQ supports Linux-based Docker containers and says Windows containers are unsupported. It also says registry pulls support public registries only. To keep a custom image private, it describes pushing the image directly to Lightsail instead; this is a different delivery path from authenticating to an existing private registry.

A team whose deployment assumes a private-registry pull needs to validate a supported image-push workflow or choose another platform. Do that compatibility check before using the low node subtotal as a buying argument. The service provides a load-balanced TLS endpoint, so do not automatically add a separate load-balancer charge merely because there are multiple nodes; verify any additional architecture requirement first.

Choose an exit policy for temporary environments

Disabling retains power, scale, deployment history and images, and the public endpoint returns HTTP503. It remains charged according to power and scale. Re-enabling restores the last active deployment and keeps the endpoint name. That can preserve a test environment, but it is not a verified cost-free weekend pause.

Deletion permanently removes the nodes, endpoint address, stored service images, deployment history and logs. The FAQ says the endpoint address cannot be recovered. Before choosing this service for throwaway previews, plan how source images, configuration, data and external routing will be retained or recreated; do not rely on deletion as an application backup.

The supplied Lightsail VM catalog contains alternatives such as a Linux/Unix IPv4 Memory optimized bundle with 2 vCPUs, 16 GB RAM and 160 GB SSD, observed on 9 November UTC. That is a VM specification, not a managed-container memory entitlement or stock reservation. Compare the VM’s operating-system work against the container service’s platform boundaries. Seasonal evidence remains in the Black Friday hub, and no offer is inferred from these ordinary tariffs.