Skip to main content
A subprocessor is a third-party data processor engaged by Coalesce who has, or potentially will have, access to or processes personal data on behalf of a Customer. This list covers the Coalesce Quality platform — the infrastructure and services that run the product and process customer data within it. Coalesce Quality is a separate product on separate infrastructure, so it has its own set of platform subprocessors; where the same vendor also appears on the Coalesce Transform list, it is because both products independently use it, not because the deployments are shared. Business and support systems — CRM, customer support, ticketing, corporate email and collaboration — are shared across Coalesce and are covered by the Coalesce subprocessor list rather than repeated here. See the Security Policy for the architecture, data access model and encryption detail behind this list.

​
Regional model

Coalesce Quality runs as independent, fully region-scoped deployments. Each region has its own Google Cloud project, its own Kubernetes cluster, its own databases, its own identity tenant and its own AI endpoints. Customer data written in a region is stored and processed only by subprocessors operating in that region. There is no cross-region replication of customer data.

​
Region-scoped subprocessors

Customer data stays within the customer’s region.

​
AI processing

All AI inference goes through Google Vertex AI on regional endpoints only — never a global endpoint. Every model is consumed as a Vertex AI model under Google’s terms, so there is no direct contract with, or direct data flow to, the model providers, and no customer data is used for model training. Requests are routed by a gateway that Coalesce operates inside the region’s own cluster.

​
Subprocessors that are not region-scoped

These process a narrow, defined slice of data and are not partitioned by region. Sentry, PostHog and Segment receive browser and web-server telemetry only. They do not receive warehouse data, query results, or data-quality results. Their scope is user identifier, workspace identifier, page and interaction events, and error diagnostics.

​
Customer-configured destinations

These receive data only where you explicitly configure the corresponding integration, and only the alert or notification content that integration is set up to deliver. See Alert integrations for how these are configured.

​
Operated by Coalesce

The following run inside each region’s own infrastructure and are not third-party subprocessors:
  • Caching and queuing — Redis, self-hosted in each region’s Kubernetes cluster.
  • Search indexing — Meilisearch, self-hosted in each region’s cluster.
  • Observability — OpenTelemetry collectors writing to the region’s own ClickHouse. No third-party observability vendor holds Coalesce Quality telemetry.
  • AI gateway — self-hosted in each region’s cluster, in front of Vertex AI.