Frequently Asked Questions
We support the following stable versions of Neo4j:
- Neo4j Community Edition 1.8.3
- Neo4j Community Edition 1.9.2
- Neo4j Community Edition 1.9.3
- Neo4j Community Edition 1.9.4
- Neo4j Community Edition 1.9.5
- Neo4j Community Edition 2.0.0
When provisioning an instance through our Heroku addon, the latest stable version will be used by default.
You can deploy a specific version by passing a
--version param to the Heroku CLI with the corresponding version string.
Example for Neo4j v 1.9.5:
$ heroku addons:add graphenedb --version v195
This is a list of the versions currently available and the corresponding version parameter strings:
We support all features of Neo4j.
Note: Since 2.x versions, Gremlin is no longer shipped with Neo4j. Read more at Using Gremlin on GrapheneDB.
We recommend that you use the Cypher query language, but you can also use Gremlin or REST traversals to query your instance.
We test some drivers explicitely to ensure that they work properly with our hosting serice. The drivers we have tested to work with GrapheneDB are:
- Java OGM
- Neocons (Clojure)
- Neography (Ruby)
- Node-neo4j (Node.js)
- Seraph (Node.js)
- Py2Neo (Python)
- Python REST Client
We will continue to work on testing more drivers to detect any compatibility problems and make sure you don't experience any problems with official drivers.
Besides our list of verified drivers, there is a whole range of official Neo4j drivers that you can use with GrapheneDB.
If you encounter any problems using any official Neo4j driver please get in touch with us. We will be glad to help you out.
Our dashboard is a rich front end application that should work with any popular modern browser. If you use our app with an older or unsupported browser, a warning message is displayed at the top of the page. Also, notice you might encounter errors while using the dashboard.
We currently support:
- Microsoft Internet Explorer 10 and higher [download]
- Safari 5 and higher [download]
- Google Chrome 21 and higher [download]
- Firefox 19 and higher [download]
- Opera 13 and higher [download]
A warning is displayed below the link to the Neo4j/Browser interface if it's not possible to open the interface through HTTPS.
There are different scenarios where this will be the case:
The plan does not support SSL. You can upgrade to one of our production plans which supports SSL. Check out our AWS plans and pricing page, our Microsoft Azure plans and pricing page or our Heroku add-on page to see which plans support SSL connections.
The plan does support SSL, but you're running an older version of Neo4j. The old webadmin does not work properly in SSL proxy scenarios (Github issue). We encourage you to upgrade to a newer version of Neo4j. Versions 2.0.x and higher work just fine.
There is a manual export feature that enables you to download a zipped file with your database. This feature is available across all plans, including Sandbox. This can be used together with the restore feature to clone your databases within GrapheneDB or to take your data elsewhere (i.e. your development machine).
Since there is no version upgrade feature, the current workaround is to create a new database and use the export + restore features to clone your original database into an instance running a newer version of Neo4j.
All production-ready plans (Standard and Performance tier) include scheduled and manual hot backups with one week retention.
Our approach is to provide plans for different DB sizes and performance needs, so you can easily scale your DB instance as your business matures. You can start using our free sandbox plan and scale as you need.
Legacy Sandbox databases have following limitations:
- 256MB memory
- 512MB physical space
- a sensible CPU limit
New sandbox databases have following limitations:
- 1000 nodes
- 10,000 relationships
- a sensible CPU limit
You can create up to 10 sandbox instances within a standard account. If you need to create more instaces for a sensible reason, please contact us and we will consider raising your limit.
Upon 3 hours of networking inactivity, sandbox instances are set into sleeping state. Upon the next request, they are awakened and the request attended as soon as the database is up and running. This wake-up process takes approximately 5 seconds, so there is no need to worry. Your app won't suffer from any DB unavailablity, just a slight delay on the first request after a sleeping period.
This is quite similar to the way Heroku dynos are set to sleep. It just takes less time to spin up our database instances.
Note: This only applies to our free sandbox instances.
Where are databases hosted?
Can I use GrapheneDB with Heroku/EngineYard/other?
We provide connection settings that you can use to connect your app from anywhere using the Neo4j REST API.
GrapheneDB allows you to select where to host your databases. To get the best network performance, we recommend you choose the same data center that hosts your application.
This is a list of PaaS providers and recommended regions:
|Provider||AWS Region to select in GrapheneDB|
|Heroku US||AWS us-east-1|
|Heroku EU||AWS eu-west-1|
|CloudBees US||AWS us-east-1|
|CloudBees EU||AWS eu-west-1|
If you’re unsure how to determine the best region, please contact us at firstname.lastname@example.org.
When a database exceeds the maximum storage size or the node or relationship maximum counts allowed for the plan, the database will be switched to read-only mode.
If you wish to turn the database back to read-write mode, the simplest solution is to increase the database capacity by upgrading to a higher plan. This can be done by cloning the database from the UI using the Upgrade/Clone button. Please note that a new instance will be created and the original instance will also be kept running. It's your responsibility to remove the original database if you don't need it anymore to avoid any potential charges for the original database.
If you don't want to upgrade into a higher plan, but would rather like to reduce the size to be within the capacity limitations of your current plan, then you can follow these steps:
- Unpack the downloaded tarball and put the directory inside the /data directory of your local Neo4j installation.
- Start your local instance and reduce your dataset
- Stop your local instance, compress the dataset and restore your GrapheneDB database with the compressed file (Admin tab, Restore database button)
With GrapheneDB, you only pay for time that your database is running. The rate for your plan is pro-rated at the end of each month. You will be billed for the number of days that your plan is in use. If you have multiple plans, a single payment is collected at the end of each month for all plans that you select.
From 1 January 2015, new EU VAT rules have come into effect changing the place of supply in respect of all supplies of telecommunications, broadcasting and e-services to consumers from the place where the supplier is located to the place where the consumer resides. These changes affect GrapheneDB and its customers.
GrapheneDB will apply the local VAT (at the place of residence of customer) for any EU customer who fails to prove itself as a company or professional.
Customers from outside the EU
Customers outside of the EU won’t be applied any VAT, regardless if they are a business, professional or a consumer. The only requisite is that we validate that the customer resides in a country outside the EU.
We will interpret a customer resides outside the EU if we find at least one evidence matching the country entered in the billing address:
- Country where credit card on file was issued
- IP address used when submitting the billing details
In case that we don’t find any evidence matching the billing address country we will ask the customer to, either repeat the process using a different card or IP address, change their billing address country, or provide any official document as proof.
European companies and professionals won’t be charged any VAT. Consumers will be applied the VAT in their country of residence, i.e. 19% for German customers, 21% for Spanish customers.
The first necessary step is to validate the country entered in the billing address as the country of residence. The validation is successful in case any of the following match the billing address country:
- Country where credit card on file was issued
- IP address used when submitting the billing details
- Country of the supplied European VAT number, provided it’s valid
If we can’t validate the country in the billing address by any of these means we will contact the customer and ask to submit again any of these or provide a document as proof.
The second and final step is only applicable to businesses. If we validate the customer as a company or professional, no VAT will be applied. The standard method for this validation is to verify the European VAT number supplied by the customer.
There is an official website which can be used to check the validity of an EU VAT number: http://ec.europa.eu/taxation_customs/vies/?locale=en
The EU VAT number can be entered in an input field for that purpose in the billing details form:
- When adding the billing details during the provisioning process or
- from the Billing tab in the Account section.
Should we fail to verify the validity of the EU VAT number provided, or in case the customer is has not registered as an intra-community VAT operator, other proof such as legal documents can be provided.
Customers can know if they have been validated as EU businesses or not by visiting the Billing tab in the Account section:
A verified EU business will have a tick icon, along with a text indicated they are verified.
For EU consumers and EU customers who have failed to verify as a business, there will be a text communicating the fact that VAT charges will apply.
Users of the Heroku GrapheneDB add-on are billed by Heroku. We don’t know if Heroku will be applying this change in the law to their European customers now or in the future. Any questions regarding Heroku billing should be directed to the Heroku support team.
Questions and requests for assistance in the verification process should be directed to our support team.
Our account deletion process is not yet automated. There is quite some complexity involved in this process from our side:
- Payments and invoices are handled automatically every month by our billing system, so our current approach is to wait until the end of the current monthly billing cycle. Once we verify that we have collected any outstanding payments and sent the corresponding invoices to the user, we then proceed to delete the account.
- This opens up an issue where you might still have paid databases running when asking us to close your account, and in that case the billable amount would keep increasing until they are deleted.
Until this process gets the attention it deserves and we implement a fully automated process with immediate checkout and payment, this is how we are handling account deletion requests:
- We ask our users to delete any paid databases on their accounts before submitting the request to delete their accounts. This will ensure that the outstanding amount does not increase due to additional billing units that might add up until we collect the payments.
- We then take them to a support ticket form where they can submit their request.
- We will wait until the next billing cycle (once a month) and collect any outstanding payments on the account.
- We will let the users know once their accounts has been successfully deleted.
We'd like to thank your being patient and wait until we have made sure everything is ok and we can close your account. If for any reason, you need to delete your account immediately, please get in touch and we will deal with it ASAP.
Support tickets can be opened from the user interface by following the support link. The support ticket form lets users select the affected database and indicate a priority. Urgent tickets on production databases (Standard and Performance plans) will be prioritized and escalated to our on-call engineers depending on the plan and urgency.
You can also reach us via email at email@example.com. Please note that issues received by email are not automatically assigned any priorities and thus, they will be treated with a low priority.