PD AWS Neptune for Blesta

PD AWS Neptune lets a hosting company sell Amazon Neptune, AWS's managed graph database, as its own package in Blesta. Your customer orders and pays, a dedicated cluster is built in your AWS account, and your customer manages it from the Blesta client portal: query endpoints, usage graphs, snapshots and an allowed-address list, with no AWS console and no mention of Amazon. You set the price, AWS bills you the cost, and the difference is your margin.

It is built for hosts whose customers work with connected data: recommendation engines, fraud and identity graphs, knowledge graphs and network maps. Those teams are comfortable with Gremlin, openCypher or SPARQL, rarely want to run a graph database themselves, and would rather buy one from their host than open an AWS account.

Why hosts choose PD AWS Neptune

A package few hosts offer. Graph workloads are real and the teams that build them have few places to buy a managed graph database from a provider they already trust. You become that place, with your pricing and your brand on every page.

All three query languages, ready to copy. The connection details show the cluster endpoint on port 8182 and the Gremlin (wss), openCypher and SPARQL (https) endpoints, each with a copy button. They are the cluster's real query endpoints.

Every paid order sets itself up. Blesta adds a new order as pending and provisions it once the invoice is paid, through its Provision Paid Pending Services task. The cluster is built in the region the customer chose, with the plan and version they picked, and nobody on your team opens the AWS console.

Nothing is shared between customers. Each cluster gets its own security group, named for the service. If the region has no default network, or your key cannot create a security group, the order is refused with the reason in your Activity log. The cluster is never put on a shared firewall.

Told plainly how access works. Neptune has no database username or password. Access is controlled by the network, so whatever the cluster's firewall lets in can read and write the graph, and the customer's page says exactly that. Customers add the addresses of their own application servers, and nothing else gets through.

Nothing left billing. Terminating a service removes the cluster, its instances, its own firewall and its manual snapshots, then confirms at AWS that they are gone. A cluster stopped for suspension is started first so AWS will delete it.

How it works

1. Connect your AWS account. Create an IAM user with the policy in our guide. Under Settings, Company, Modules, open PD AWS Neptune and add an account with its keys. The secret is stored encrypted. Test Connection in the PD Modules panel makes one read-only call; nothing is created.
2. Create your package. Choose PD AWS Neptune as the module and pick the account, region, plan and Neptune version. The package form offers dropdowns of what is actually on offer, and private network access is the default.
3. Generate and price the options. One click in the Products panel attaches Region, Plan and Version as package options, trimmed to what AWS offers in that region, each at price zero. You set every price under Package Options. The module never changes a price after generation unless you press Rebuild, and the button warns you first.
4. Your customer orders. They choose their options and pay. Their endpoints appear on the service page as soon as AWS reports the cluster available.

What you get in Blesta

Selling and pricing

• Region, Plan and Version as package options, with plans trimmed to what AWS actually offers in the package's region
• A default plan that is always one AWS currently offers
• A blank version means whatever AWS uses for new clusters; a chosen version is the one provisioned
• A chosen option's price is what the invoice and every renewal charge, in the customer's currency
• A resource name prefix of your choice, so clusters read as <prefix>-<service id> in AWS
• Switches for each customer feature, per package: snapshots, firewall, connection details and usage graphs

Provisioning you can trust

• If AWS refuses an order, you see AWS's reason in the Activity log and the customer is told it could not be set up, never left on "still provisioning"
• Creating the service again adopts the cluster that already exists instead of building a second one
• Plan upgrades and downgrades resize the cluster in place and keep the data
• Suspending stops the cluster and keeps the data. AWS restarts a stopped cluster after seven days, and the module stops it again
• A cluster deleted outside Blesta is reported to you in the Activity log, never to the customer
• The module refuses to be deactivated while it still tracks clusters, so no record is orphaned

Running it day to day

• Clients and Databases: every cluster with its service, plan, region, endpoint and state
• Activity: recent provisioning and actions across all clients, with the outcome and AWS's reason for any failure
• A Report showing licence state, your configured AWS accounts and the database count
• Read-only AI tools for service status, the database list and recent activity; nothing can create, resize or delete through them

What your customers get

• Start, stop and reboot, each behind a confirmation dialog in the page, and pressing one twice while the first is queued does it once
• Query endpoints for Gremlin, openCypher and SPARQL, with copy buttons and a plain note on how access works
• Usage graphs for CPU, requests per second, freeable memory and storage used over the last 24 hours
• Snapshots they take in one click and delete when finished, each with its date. AWS also keeps automated daily backups for seven days
• A firewall where they add and remove IPv4 addresses or ranges. A new cluster allows nothing until they do, and an invalid entry is refused with a message
• Activity, the last 40 actions in plain words with when and how each ended

Each cluster has one instance, so a failure means a short outage rather than a failover, and the pages say so. Storage grows by itself, so there is no disk size to choose. Restoring a snapshot is your job in AWS. Clusters are private to your AWS network and never get a public address, so customers connect from an application server inside it. Every feature is a switch on the package: turn one off and it disappears from the customer's menu and the server refuses it too. The pages follow the client portal's own theme and are fully translatable, with overrides that survive updates.

Security built in

• Your AWS key is stored encrypted on the module account and never reaches a customer page.
• Neptune queries travel over TLS, and only the addresses a customer has allowed can reach the cluster.
• A suspended, terminated, cancelled or fraud-held service cannot start, snapshot or change anything.
• Only licence data reaches us. Customer data never does.

One product, every feature

PD AWS Neptune is one paid module with everything included: provisioning, power, snapshots, firewall, usage graphs and teardown. There are no tiers to compare. It installs into PD Modules in one click, updates itself, and uninstalls cleanly without modifying any Blesta files.

If your subscription ever lapses, nothing the module built is touched. You get a 14-day grace period with everything working, then admin management locks while customers keep being served, and only after that do customer features pause, behind a neutral message that never mentions licensing. Renewing restores everything at once.

Requirements

• Blesta 6 or later on PHP 8.2 or later, with the ionCube Loader
• An AWS account and an IAM user with the policy we provide
• Blesta's cron running, so paid orders are provisioned
• PD Modules, the free control panel every PD module installs into