Mason

Policies · Last updated 16 November 2026

Data Processing Addendum

How we process the data you sync into a Mason replica, as your processor: what we may do with it, who else touches it, where it lives, how it is deleted, and the security measures we keep.

Scope of this DPA

This Data Processing Addendum (“DPA”) is between TMD Technology Limited, trading as Embeddable, the company that makes Mason, registered in England and Wales under number 13856879, whose registered office is at International House, 142 Cromwell Road, London SW7 4EF (“we”, “us”, “our”), and the customer that has agreed to the Agreement (defined below) with us (“Customer”). This DPA applies to Customer automatically under the Agreement when Customer uses the Service to process Personal Data, and does not need to be signed. If Customer has entered into a separate, signed data processing agreement with us, that agreement prevails over this DPA.

This DPA governs our processing of Customer Data in the course of providing the Service under the Mason Terms of Service (masondata.ai/policies/terms) or any other agreement between Customer and us governing Customer’s use of the Service (the “Agreement”). If this DPA conflicts with the Agreement, this DPA controls. If the Standard Contractual Clauses described in Section 7 conflict with this DPA or the Agreement, the Standard Contractual Clauses control. Capitalised terms not defined in this DPA have the meaning given in the Agreement.

“Service” means Mason Cloud, the hosted service at app.masondata.ai, together with the replicas it runs for Customer and the hosted commands of the mason command-line tool. It does not include the open-source Mason software when Customer runs that software itself, which processes nothing on our systems.

Compliance with Data Protection Laws

We and Customer each agree to comply with our respective obligations under applicable data privacy and data protection laws (together, “Data Protection Laws”) in connection with the Service. Depending on the circumstances, Data Protection Laws may include:

Roles of the parties

In connection with the Agreement, Customer determines the purposes and means of the processing of Customer Data and acts as a “Data Controller”, and we process Customer Data on behalf of Customer and in accordance with Customer’s instructions as a “Data Processor”. “Data Controller” and “Data Processor” have the meanings given to those terms under applicable Data Protection Laws. Where Customer itself acts as a processor on behalf of its own customers, we act as Customer’s sub-processor.

For the purposes of the Agreement and this DPA:

(i) “Personal Data” means any information relating to an identified or identifiable natural person, as defined under applicable Data Protection Laws.

(ii) “Customer Data” means Personal Data that we process on Customer’s behalf to provide the Service, which is the Personal Data in:

Customer decides what the synced copy contains, so Customer Data may include, by way of example, identifiers of Customer’s own users and customers (such as user IDs, names and email addresses), their contact details, account, transaction and application records, and usage data, to the extent Customer’s database holds them in the tables it chooses. Customer Data does not include data that we process for our own purposes as a controller, including Service Data.

(iii) “Service Data” means Personal Data relating to Customer’s administrators, developers and other representatives who hold accounts for the Service (but not Customer’s end users) that we process to operate, secure, support and improve the Service, and for which we determine the purposes and means of processing. We process Service Data as a Data Controller, and our Privacy Policy governs that processing.

(iv) “Subprocessor” and “Subprocessor List” have the meanings given in Section 1(h).

1. Processing requirements

As a Data Processor, we agree to:

a. process Customer Data only (i) on Customer’s behalf to provide and support the Service for Customer: syncing the tables and columns Customer chooses into the replica and keeping that copy up to date, answering read-only queries from the users, tools and addresses Customer allows, logging the statements answered, investigating failures and slow statements, and monitoring the Service for abuse, security and safety; (ii) on Customer’s documented instructions, including with regard to international transfers, which are the Agreement (including the Terms of Service), this DPA and Customer’s configuration of the Service (the tables and columns it syncs, the database users, roles and addresses it allows, and the projects it deploys), together with any further instructions that we agree in writing; and (iii) in a manner that provides no less than the level of privacy protection required of us by Data Protection Laws;

b. promptly inform Customer in writing if we cannot comply with the requirements of this DPA;

c. not provide Customer with remuneration in exchange for Customer Data. The parties acknowledge and agree that Customer has not “sold” (as defined by the CCPA) Customer Data to us;

d. not “sell” (as defined by U.S. Privacy Laws) or “share” (as defined by the CCPA) Customer Data;

e. inform Customer promptly if, in our opinion, an instruction from Customer violates applicable Data Protection Laws;

f. require (i) persons employed by us and (ii) other persons engaged to perform on our behalf to be subject to a duty of confidentiality with respect to the Customer Data and to comply with the data protection obligations applicable to us under the Agreement and this DPA;

g. not use Customer Data for our own purposes. The tools, clients and AI agents that Customer connects to its replica (a BI tool, a SQL client, an MCP client, a coding agent, or any other service using Customer’s own account with that provider) are provided or engaged by Customer, and what they read from the replica is Customer’s own use of Customer Data. They are not our Subprocessors;

h. engage the organisations listed at masondata.ai/policies/subprocessors to process Customer Data (each a “Subprocessor”, and the list at that URL, the “Subprocessor List”) to help us satisfy our obligations in accordance with this DPA, or to delegate all or part of the processing activities to such Subprocessors. Customer hereby consents to the use of those Subprocessors. We will inform Customer, by email to the address of Customer’s account, of any intended addition or replacement of a Subprocessor before that Subprocessor starts to process Customer Data. If Customer does not wish to consent to the use of an additional Subprocessor, Customer may, on reasonable grounds relating to the protection of Customer Data, notify us that it does not consent by emailing harry@embeddable.com within fifteen (15) days after we inform Customer of the change. In that case, we shall have the right to cure the objection through one of the following options: (i) we will cancel our plans to use the Subprocessor with regard to processing Customer Data, or will offer an alternative way to provide the Service without that Subprocessor; (ii) we will take the corrective steps requested by Customer in its objection notice and proceed to use the Subprocessor; (iii) we may cease to provide, or Customer may agree not to use, whether temporarily or permanently, the particular aspect or feature of the Service that would involve the use of that Subprocessor; or (iv) Customer may cease providing Customer Data to us for processing involving that Subprocessor. If none of these options is commercially feasible, in our reasonable judgment, and the objection has not been resolved to the satisfaction of the parties within thirty (30) days of our receipt of Customer’s objection notice, either party may terminate, for cause, any subscriptions or usage of the Service that cannot be provided without the new Subprocessor. In that case Customer will be refunded any pre-paid fees for those subscriptions or usage to the extent they cover periods after the date of termination. This termination right is Customer’s sole and exclusive remedy if Customer objects to a new Subprocessor. We shall enter into a written agreement with each Subprocessor that imposes data protection obligations no less protective than those in this DPA, and we remain responsible to Customer for each Subprocessor’s performance of those obligations;

i. upon reasonable request, no more than once per year, provide Customer with our privacy and security policies and other information necessary to demonstrate compliance with the obligations set out in this DPA and applicable Data Protection Laws;

j. where required by law, and upon reasonable notice and appropriate confidentiality agreements, cooperate with assessments, audits or other steps performed by or on behalf of Customer, at Customer’s sole expense and in a manner that is minimally disruptive to our business, that are necessary to confirm that we are processing Customer Data in a manner consistent with this DPA. Where permitted by law, we may instead make available to Customer a summary of the results of a third-party audit or certification report relevant to our compliance with this DPA. Such results, and the results of any assessments, audits or other steps, are our Confidential Information;

k. to the extent that Customer permits or instructs us to process Customer Data subject to U.S. Privacy Laws in a deidentified, anonymised or aggregated form as part of the Service, (i) adopt reasonable measures to prevent such deidentified data from being used to infer information about, or otherwise being linked to, a particular natural person or household; (ii) not attempt to reidentify the information, except that we may attempt to reidentify the information solely to determine whether our deidentification processes comply with Data Protection Laws or are functioning as intended; and (iii) before sharing deidentified data with any other party, including Subprocessors, contractually obligate any such recipient to comply with the requirements of this provision;

l. where the Customer Data is subject to the CCPA, not (i) retain, use, disclose or otherwise process Customer Data except as necessary for the business purposes set out in Section 1(a)(i) and Exhibit A, which are the limited and specified business purposes for which Customer discloses Customer Data to us; (ii) retain, use, disclose or otherwise process Customer Data in any manner outside of the direct business relationship between us and Customer; or (iii) combine any Customer Data with Personal Data that we receive from or on behalf of any other third party or collect from our own interactions with individuals, except that we may combine Customer Data for a purpose permitted under the CCPA if directed to do so by Customer or as otherwise permitted by the CCPA; and

m. where required by law, grant Customer the rights to (i) take reasonable and appropriate steps to ensure that we use Customer Data in a manner consistent with Data Protection Laws by exercising the audit provisions in Section 1(j); and (ii) stop and remediate unauthorised use of Customer Data, for example by requesting that we provide written confirmation that applicable Customer Data has been deleted.

2. Notice to Customer

We will inform Customer if we become aware of:

a. any legally binding request for disclosure of Customer Data by a law enforcement authority, unless we are forbidden by law to inform Customer, for example to preserve the confidentiality of a law enforcement investigation;

b. any notice, inquiry or investigation by a supervisory authority established under Article 51 of the GDPR (a “Supervisory Authority”) with respect to Customer Data; or

c. any complaint or request (in particular, requests for access to, rectification or blocking of Customer Data) received directly from Customer’s data subjects. We will not respond to any such request without Customer’s prior written authorisation.

3. Assistance to Customer

We will provide reasonable assistance to Customer regarding:

a. information necessary, taking into account the nature of the processing, to respond to requests received under Data Protection Laws from Customer’s data subjects in respect of access to, or the rectification, erasure, restriction, portability, objection, blocking or deletion of, Customer Data that we process for Customer. Because the synced copy follows Customer’s database, a correction or erasure that Customer makes in its database reaches the replica at its next sync. If a data subject sends such a request directly to us, we will promptly send the request to Customer;

b. the investigation of any breach of the security of us or our Subprocessors leading to the accidental or unlawful destruction, loss, alteration, unauthorised disclosure of, or unauthorised access to, Customer Data (a “Personal Data Breach”), and Customer’s notification of a Personal Data Breach to a Supervisory Authority or to data subjects, taking into account the nature of the processing and the information available to us; and

c. where doing so is mandatory, the preparation of data protection impact assessments with respect to the processing of Customer Data by us and, where necessary, consultations with any Supervisory Authority with jurisdiction over that processing.

4. Required processing

If we are required by law to process any Customer Data for a reason other than in connection with the Agreement, we will inform Customer of this requirement before any such processing, unless legally prohibited.

5. Security

We will:

a. maintain reasonable and appropriate organisational and technical security measures (including with respect to personnel, facilities, hardware and software, storage and networks, access controls, monitoring and logging, vulnerability and breach detection, incident response, and encryption), including those described in Exhibit B, to protect against unauthorised or accidental access, loss, alteration, disclosure or destruction of Customer Data and to protect the rights of the subjects of that Customer Data;

b. take appropriate steps to confirm that our personnel are protecting the security, privacy and confidentiality of Customer Data consistent with the requirements of this DPA; and

c. notify Customer of any Personal Data Breach by us, our Subprocessors or any other third parties acting on our behalf without undue delay and, where feasible, within seventy-two (72) hours after we become aware of it.

6. Obligations of Customer

a. Customer represents, warrants and covenants that it has and shall maintain throughout the term all necessary rights, consents and authorisations to provide the Customer Data to us and to authorise us to use, disclose, retain and otherwise process Customer Data as contemplated by this DPA, the Agreement and any other processing instructions provided to us.

b. Customer shall comply with all applicable Data Protection Laws.

c. Customer shall reasonably cooperate with us to assist us in performing any of our obligations with regard to requests from Customer’s data subjects.

d. Without prejudice to our security obligations in Section 5, Customer acknowledges and agrees that it, rather than us, is responsible for certain configurations and design decisions for the Service, and that Customer, not us, is responsible for implementing those configurations and design decisions in a secure manner that complies with applicable Data Protection Laws. In particular, Customer chooses which tables and columns are synced, which database users exist on the replica and which role each has, which addresses each user may connect from, and which tools and agents it connects.

e. Customer shall not provide Customer Data to us except through agreed mechanisms. Without limiting the foregoing, Customer represents, warrants and covenants that it shall only transfer Customer Data to us using secure, reasonable and appropriate mechanisms, to the extent such mechanisms are within Customer’s control.

f. Customer shall not take any action that would (i) render the provision of Customer Data to us a “sale” under U.S. Privacy Laws or a “share” under the CCPA (or equivalent concepts under U.S. Privacy Laws); or (ii) render us not a “service provider” under the CCPA or a “processor” under U.S. Privacy Laws.

g. Customer controls which data the Service can access and is responsible for the Customer Data it chooses to sync. The Service lets Customer leave out any column of a table. Customer shall not sync into a replica: payment card numbers; bank account numbers; government identification numbers (such as passport, driving licence, national insurance or social security numbers); passwords, password hashes or other login credentials of individuals; or special categories of personal data (as defined in Article 9 of the GDPR) or personal data relating to criminal convictions and offences.

h. Customer’s own database remains Customer’s system of record. The synced copy on a replica is derived from it and can be recreated from it. Customer is responsible for backing up its own database.

7. International transfers and Standard Contractual Clauses

a. We are established in the United Kingdom. We host Customer Data with Amazon Web Services in its eu-west-1 region in Ireland, and we and our Subprocessors may process Customer Data in the locations shown in the Subprocessor List. Transfers of Customer Data from the European Economic Area (“EEA”) to us in the United Kingdom are covered by the European Commission’s adequacy decision for the United Kingdom, and transfers from Switzerland are covered by Switzerland’s recognition of the United Kingdom as providing adequate protection. Sections 7(b) to 7(d) apply only to the extent that a transfer of Customer Data to us is a restricted transfer under Data Protection Laws that is not covered by an adequacy decision or another lawful transfer mechanism.

b. EEA transfers. The standard contractual clauses adopted by the European Commission in Implementing Decision (EU) 2021/914 of 4 June 2021 (the “EU SCCs”) are deemed entered into, are incorporated into this DPA by reference, and are completed as follows:

  1. Module Two (controller to processor) applies where Customer is a controller, and Module Three (processor to processor) applies where Customer is a processor;
  2. Clause 7 (docking clause) does not apply;
  3. in Clause 9(a), Option 2 (general written authorisation) applies, and the time period for prior notice of changes to Subprocessors is fifteen (15) days;
  4. the optional wording in Clause 11(a) does not apply;
  5. in Clause 13, the competent supervisory authority is as set out in Part C of Exhibit A;
  6. in Clause 17, Option 1 applies, and the EU SCCs are governed by the law of Ireland;
  7. in Clause 18(b), disputes are resolved by the courts of Ireland; and
  8. Annex I is completed with the information in Exhibit A, Annex II with the information in Exhibit B, and the list of sub-processors is the Subprocessor List.

c. Swiss transfers. Where the FADP applies to a transfer, the EU SCCs apply as completed in Section 7(b), with these changes: (i) references to the GDPR are to be read as references to the FADP; (ii) the Swiss Federal Data Protection and Information Commissioner is the competent supervisory authority for transfers governed by the FADP; and (iii) the term “Member State” in Clause 18(c) must not be interpreted in a way that excludes data subjects in Switzerland from bringing claims in their place of habitual residence (Switzerland).

d. UK transfers. Where the UK GDPR applies to a transfer, the parties will comply with Part 2: Mandatory Clauses of the International Data Transfer Addendum to the EU SCCs, being template Addendum B.1.0 issued by the Information Commissioner’s Office and laid before Parliament in accordance with s119A of the Data Protection Act 2018 on 2 February 2022, as revised under Section 18 of those Mandatory Clauses (the “UK Addendum”). For Part 1 of the UK Addendum: Table 1 is completed with the information in Exhibit A; Table 2 refers to the EU SCCs as completed in Section 7(b); Table 3 is completed with Exhibit A, Exhibit B and the Subprocessor List; and for Table 4, either party may end the UK Addendum as set out in Section 19 of the UK Addendum.

e. In this DPA, “Standard Contractual Clauses” means the EU SCCs as completed under Section 7(b), including as applied under Sections 7(c) and 7(d).

8. Term, return and deletion

This DPA remains in effect for as long as we process Customer Data on Customer’s behalf, and in any case until the Agreement ends and all Customer Data has been deleted in accordance with this Section 8.

Customer’s database stays with Customer. The synced copy on a replica is a copy of data that Customer already holds, so there is nothing for us to return: Customer’s own database is the original.

When Customer deletes a replica, or when the Agreement ends, we destroy the replica’s machine, its encrypted volume (which holds the synced copy and the replica’s log of statements) and the secret that held Customer’s database credentials. We keep no backup or snapshot of a replica. Deletion also drops the replication slot that the replica held on Customer’s database, and the Service checks that it is gone, or tells Customer how to drop it by hand.

The incident and diagnostic records described under “Roles of the parties” expire fourteen (14) days after each was written, so they can outlive a replica by up to fourteen days. No Customer Data is retained beyond these periods unless applicable law requires it.

When the Agreement ends, we will also direct each Subprocessor that holds Customer Data to delete it, unless the law prohibits this.

Exhibit A: Description of the processing

This Exhibit describes the processing of Customer Data under this DPA. It is also Annex I to the Standard Contractual Clauses and provides the Table 1 and Table 3 information for the UK Addendum.

A. List of parties

Data exporterData importer
NameCustomer, as identified in Customer’s account with usTMD Technology Limited (trading as Embeddable), company no. 13856879
AddressAs identified in Customer’s account with usInternational House, 142 Cromwell Road, London SW7 4EF, United Kingdom
Contact personAs identified in Customer’s account with usHenry Marshall, COO, harry@embeddable.com
Activities relevant to the transferUse of the Service under the AgreementProvision of the Service under the Agreement
RoleController, or processor where Customer acts on behalf of its own customersProcessor, or sub-processor where Customer is a processor
Signature and dateCustomer’s acceptance of the Agreement is its signature of the Standard Contractual Clauses and this Exhibit, dated the date Customer accepted the AgreementOur entry into the Agreement is our signature of the Standard Contractual Clauses and this Exhibit, dated the date Customer accepted the Agreement

B. Description of the processing and transfer

Categories of data subjects. Customer determines and controls the Personal Data it syncs into a replica, which may relate to:

  • Customer’s own customers and end users, and the users of Customer’s customers;
  • employees, consultants, contractors and agents of Customer;
  • suppliers, partners and other individuals with whom Customer does business; and
  • any other individual whose Personal Data is held in the tables Customer chooses to sync.

Categories of Personal Data. Whatever the tables and columns Customer chooses to sync contain, which may include identifiers (such as user IDs, names and email addresses), contact details, account, transaction and application records, and usage data; together with the names of the database users Customer creates on the replica, as they appear in the replica’s log of statements.

Sensitive data. None. Under Section 6(g), Customer does not sync special categories of personal data, personal data relating to criminal convictions and offences, or the other data listed there.

Frequency of the transfer. Continuous, for the term of the Agreement: on demand, by polling or by logical replication, as Customer configures each table.

Nature of the processing. Connecting to Customer’s database with read-only access; copying the tables and columns Customer chooses onto a dedicated, encrypted volume in the eu-west-1 region and keeping the copy up to date; answering read-only SQL and MCP queries over the copy from the database users, tools and addresses Customer allows; logging the statements answered; recording failed and slow statements, error messages and table structures for diagnostics; and providing support at Customer’s request.

Purpose of the processing. To provide the Service to Customer, as described in Section 1(a)(i) and under “Nature of the processing” above.

Retention. The synced copy and the replica’s log of statements (the log’s default is seven days): until Customer deletes the replica. Incident and diagnostic records: fourteen days. Everything else: as set out in Section 8.

Transfers to Subprocessors. To the Subprocessors in the Subprocessor List, for the purposes and in the locations stated there, for the term of the Agreement.

C. Competent supervisory authority

For the EU SCCs, under Clause 13: where Customer is established in an EEA Member State, the supervisory authority responsible for ensuring Customer’s compliance with the EU GDPR for the transfer; where Customer is not established in the EEA but has appointed a representative under Article 27(1) of the EU GDPR, the supervisory authority of the Member State in which that representative is established; and otherwise the supervisory authority of the Member State in which most of the data subjects whose Personal Data is transferred are located. For transfers under the UK Addendum, the Information Commissioner’s Office. For transfers governed by the FADP, the Swiss Federal Data Protection and Information Commissioner.

Exhibit B: Technical and organisational security measures

This Exhibit is Annex II to the Standard Contractual Clauses. It describes the security measures we maintain for the Service and for Customer Data. The Service is in beta, and these measures are what it has today; we will update this Exhibit as they change.

Independent assessment

  • We do not yet hold a SOC 2 report or an ISO 27001 certification for the Service. Amazon Web Services’ own compliance documentation covers the infrastructure the Service runs on.

Where Customer Data lives

  • Every replica, the control plane that manages them, its database, its logs and its secrets run in Amazon Web Services’ eu-west-1 region in Ireland. Customer Data leaves that region only as the Subprocessor List describes.
  • Each replica runs on a virtual machine of its own, with its own encrypted volume. No two customers share a machine or a volume.

Isolation and network access

  • A replica’s machine sits in a private subnet with no public IP address, accepts no SSH, requires IMDSv2 for its own credentials, and is reachable only through its own load balancer on port 5432.
  • A connection reaches a replica only from an address that one of its active database users is allowed to connect from, or from anywhere while Customer has made a user public. With no users, nothing reaches the replica. On a replica created since users gained their own addresses, each user’s addresses are enforced twice: by the load balancer’s security group and by Mason itself. An older replica has one list of addresses, enforced by its security group.
  • Every connection to a replica uses TLS, with a certificate from a publicly trusted authority, and a generated password, which is shown to Customer once and never kept by the control plane.
  • Replicas reach Customer’s database through a NAT gateway with fixed outbound IP addresses, which the Service shows Customer before a replica exists, so that Customer can allow those addresses alone on its database’s firewall.

Customer’s database credentials

  • Credentials are entered in one place: the replica creation form, or the mason replica create command, which reads them from a hidden prompt or standard input and never from a flag or a file. The control plane holds them in memory only, on their way to Mason’s checks of the database and to the replica’s secret.
  • The secret is held in AWS Secrets Manager. The control plane can write it but has no permission to read it back; only the replica’s own machine reads it, with its own instance role. The secret is destroyed with the replica, without a recovery window.
  • Where Customer’s database is reached through an SSH bastion, the Service generates a key pair for the replica; the private key is sealed with AES-256-GCM on its way to the replica’s secret and is kept nowhere else.
  • Mason’s checks of a database run in an isolated function with the credentials of that one run, which writes only its own logs; a later check of a database reached through Customer’s SSH bastion runs on the replica’s own machine instead. Failed checks are limited per person and per database user, with a daily cap across every database, so the checks cannot be used to guess a password.
  • The control plane never stores or logs a credential. Its audit log refuses to write any value shaped like a secret, and its logs name errors by their class alone.

Read-only on Customer’s database

  • Mason only reads from Customer’s database. The Service shows Customer the SQL to create a read-only database user for Mason, and warns at creation if the user it was given could write.
  • Mason connects to Customer’s database with TLS (sslmode=require). It does not verify the database’s certificate, and has no setting for a private certificate authority.

Encryption

  • At rest: each replica’s volume is encrypted (AWS EBS encryption); the control plane’s database is encrypted and enforces TLS; incident records are kept in a log group encrypted with a key of its own, whose use is recorded by AWS CloudTrail.
  • In transit: TLS between Customer’s tools and a replica, between the mason command and the control plane, between the control plane and each replica’s machine, and between a replica and Customer’s database.

Tokens, sessions and the web application

  • Login tokens, CI tokens and host-agent tokens are stored as SHA-256 hashes. A new token is shown once, in the answer to the request that made it, and never logged.
  • Sessions are signed cookies that expire after 12 hours; signing out revokes the session so that a copied cookie stops working.
  • The web application runs scripts only from its own origin, with no inline script, and every page works without script. A replica’s own page is relayed inside a sandbox with a content security policy of its own.

Our team’s access to Customer Data

  • No member of our team can log in to a replica’s machine: it accepts no SSH and has no public address.
  • Incident records can be decrypted only by a named operator signing in to AWS with single sign-on, and by our diagnostics workflow. The workflow runs under a read-only role that only that workflow, on the main branch of our repository, may assume; it reports a replica’s incidents as shapes with every table, column, value and statement removed, and prints the records themselves only encrypted to a key that the person investigating generates for that one session. Each decryption of the records is recorded by AWS CloudTrail.
  • Access to the control plane’s database and to AWS is given to team members on a need-to-use basis, through single sign-on with two-step verification, and can be revoked at any time.
  • We do not access Customer Data to diagnose a problem except through the incident records above, or with Customer’s written permission, or where needed to investigate or respond to a security incident or to comply with the law.

Deployment and change control

  • Only our continuous-integration system deploys the control plane, from the main branch of our repository, using short-lived credentials obtained through OpenID Connect. No long-lived cloud keys exist, and no one deploys from their own machine.
  • Each replica is defined by a versioned template, and runs a Mason image pinned by its digest. A replica changes what it runs only when Customer upgrades it, or when we apply an upgrade we have announced.
  • Every change to the Service is reviewed and tested before it is merged.

Deletion

  • Deleting a replica drops its replication slot on Customer’s database, then destroys its machine, its volume, its load balancer and its secret. The volume is deleted with the machine and is never snapshotted.
  • Incident records expire after fourteen days. The control plane’s logs expire after one month, and its database backups after one day.

Monitoring and incident response

  • Each replica reports its health to the control plane once a minute. Alarms on failures, resource use and data transfer reach our team by email within minutes, naming the replica by its id alone.
  • Service logs are kept centrally for one month and hold no Customer Data.
  • We notify Customer of a Personal Data Breach as Section 5(c) provides, and accept vulnerability reports as described in the Mason repository’s SECURITY.md.

Our team and its devices

  • Team members sign in to the systems we use with Google Sign-in and must use two-step verification. Account sharing is not allowed. We receive alerts for sign-ins from new devices or locations, and we can revoke any team member’s access at any time, for example when they leave.
  • All company hardware, and any personal hardware used for company purposes, must be encrypted at rest and protected by up-to-date anti-malware software, secured by a password and locked when unattended. Old devices are securely erased before disposal.
  • Everyone who works for us is bound by a duty of confidentiality.