Public IPv4 bundle
Use when users, upstream services, allow lists, administration or monitoring still require public IPv4. A default public address may change after stop/start; attach a static IP when stable addressing is required.
Simplified AWS infrastructure decision guide
Lightsail combines a virtual server, SSD storage and transfer allowance in a predictable bundle. The defensible buying question is whether the address family, region, recovery method, operating ownership and residual resources fit the workload after launch—not whether a blueprint can deploy it quickly.
Commercial and technical evidence checked 1 October 2026. Prices and availability can change. No private benchmark, uptime test, support-ticket test, user survey or paid ranking is claimed.
Start with ownership
Lightsail is self-managed infrastructure that lowers provisioning friction through bundles, a console, networking controls, metrics and application images. Those conveniences do not transfer day-two responsibility for Linux, WordPress, databases, access control, certificates, backups or incident response to AWS.
| Layer | Primary owner | Evidence before production |
|---|---|---|
| AWS infrastructure and Lightsail service | AWS | Current service documentation, status and support route |
| AWS account, IAM and credentials | Customer | MFA, least privilege, recovery contacts and billing alerts |
| Operating system and packages | Customer | Patch cadence, SSH controls, firewall review and logs |
| Application, plugins and database | Customer | Update ownership, data-integrity controls and incident runbook |
| Backups and restore proof | Customer | Application-aware copy, off-platform copy and isolated restore |
| DNS, certificates and integrations | Customer or selected vendor | Expiry monitoring, dependency inventory and rollback path |
Management boundary: treat a preconfigured WordPress image as a starting state, not a managed WordPress contract. If nobody owns patching, monitoring and recovery, compare a managed application platform or another service with an explicit operating scope.
Bundle-fit gate
As of 1 October 2026, AWS lists separate public-IPv4 and IPv6-only server bundles. A representative Linux general-purpose bundle is USD 5 per month with public IPv4 and USD 3.50 per month with IPv6 only. Those examples are not a universal starting cost: operating system, plan family, region and dependencies can change the decision.
Use when users, upstream services, allow lists, administration or monitoring still require public IPv4. A default public address may change after stop/start; attach a static IP when stable addressing is required.
Consider only after testing every public and administrative path. A lower bundle price does not compensate for an unreachable dependency or an emergency access path that exists only on paper.
AWS currently states that affected regions receive half of the listed transfer allowance: Mumbai, Sydney, Jakarta, Malaysia, Hong Kong and São Paulo. Recheck the exact region before publication or purchase.
Complete cost
A useful estimate includes the server, snapshot storage, attached disks, static addresses, databases, load balancers, object storage, distributions, DNS usage, tax, support and operator time. A stopped instance remains billable; stopping it does not stop its charge.
| Item | Question | Failure if omitted |
|---|---|---|
| Instance bundle | OS, address family, family, size and exact region? | False starting-price comparison |
| Transfer | How much inbound plus outbound consumes the allowance, and what excess outbound remains? | Overage or premature upsizing |
| Snapshots and disks | What persists independently and is billed by stored capacity? | Residual cost after deletion |
| Network services | Static IP, load balancer, CDN/distribution and DNS use? | Hidden architecture cost |
| Data services | Managed database, object storage, backup and export requirements? | Incomplete recovery and exit budget |
| Operations | Patching, monitoring, incident response and restore-test hours? | Cheap infrastructure, expensive unmanaged risk |
Enter monthly estimates in gigabytes. This models allowance consumption only; it does not estimate the current regional overage price.
Finish both decision gates before opening the comparison workspace. This keeps an undefined network or transfer assumption from becoming a false shortlist.
Your selections stay in this browser and are neither stored nor transmitted.
Transfer rule: AWS states that both inbound and outbound traffic count toward the allowance, while only outbound traffic above the allowance is billed as overage. Model both directions; price only the calculated excess outbound using the current regional rate.
Recovery proof
AWS documentation describes seven automatic daily snapshots as rolling copies, retaining the seven most recent snapshots. Manual snapshots persist until deletion and storage is billed separately. A crash-consistent disk image may still capture a database or application in an unusable state.
Convenient rolling retention. Confirm enablement, timing, resource coverage, storage cost and what happens when resources are deleted.
Use database dumps, platform exports or an application-supported backup method so data consistency is intentional rather than accidental.
Keep a portable copy outside the Lightsail failure, account and deletion boundary. Document credentials and restoration dependencies separately.
Complete all gates before treating the workload as recoverable or safe to delete.
Scale and exit boundary
Scaling a server normally means creating a new larger instance from a snapshot. AWS also documents export of supported manual instance and block-storage snapshots to EC2. Network, data and application services still require separate recreation, credentials, validation and cutover work.
| Resource | Snapshot/export path | Required work |
|---|---|---|
| Supported instance snapshot | Manual snapshot can be exported to EC2 | Verify compatibility, resulting AMI, instance type, networking, boot and application behavior |
| Supported block-storage snapshot | Manual disk snapshot can be exported | Reattach, mount, validate permissions and prove application consistency |
| Static IP and firewall | Not carried as the same Lightsail resource | Design EC2 networking, security groups, addressing and allow-list changes |
| Load balancer and DNS | Separate recreation | Rebuild listeners, health checks, certificates, records and rollback |
| Managed database | Separate data migration | Choose target, export/import or replicate, test consistency and manage cutover |
| Object storage and distribution | Separate inventory and migration | Move data/configuration, permissions, origins, caching and invalidation rules |
Decision trigger: evaluate EC2 or another platform before constraints force an emergency migration—especially when architecture needs exceed Lightsail’s simplified networking, automation, scaling or service boundaries.
Residual-cost control
Use the migration acceptance process · Design independent recovery · Model the full cost
Alternatives
Compare Droplet responsibility, regional service coverage, backup boundaries and complete monthly cost.
Compare exact plan/region fit, stopped-instance billing, backup retention and exit controls.
Compare network-zone pricing, processor architecture, separate IP resources and Volume recovery.
Compare compute duty cycle, core versus distributed regions, transfer pooling and backup exclusions.
Use the support-scope guide, location method, payment checks, comparator, decision tools and migration path for the wider decision.
Evidence trail
AWS facts and dated commercial examples are drawn from current AWS pages. Direct, untracked AWS links preserve the commercial boundary.
No SLA percentage is published here: the current legal scope, exclusions and claim procedure were not independently verified to the standard required for a consequential uptime claim.
Questions
No. Lightsail simplifies provisioning, bundles and monitoring, but the customer remains responsible for operating-system and application maintenance, security, data protection and incident response.
No. A stopped Lightsail instance continues to incur charges because its resources remain allocated. Delete resources only after migration and recovery evidence is complete.
IPv6-only bundles omit a public IPv4 address and currently cost less. Use one only after proving that users, dependencies, administration and monitoring work over IPv6 or through an appropriate intermediary.
Both inbound and outbound data count toward the included allowance. Only outbound transfer above the allowance is billed as overage, subject to current regional rules.
AWS documentation states that automatic snapshots retain the seven most recent daily snapshots. They are not a substitute for application-aware and off-platform recovery copies.
Supported manual instance and disk snapshots can be exported to Amazon EC2. Other resources require separate inventory, recreation and testing.
Not necessarily. Snapshots, detached static IPs, disks, databases, load balancers, object storage, distributions, DNS and other resources require a separate residual-resource sweep.
No verified AWS or Amazon Lightsail affiliate relationship was identified. AWS links on this page are direct and untracked.
If your decision turns on custom networking, policy-driven scaling, specialized compute or a future exit, use the Amazon Lightsail vs EC2 workload, cost and migration guide. It adds a whole-architecture inventory and a reversible migration gate without duplicating this Lightsail dossier.