Architecture & data flow

High-level view of the managed Evitrus cloud on evitrus.io for due diligence.

Component diagram

flowchart LR
  subgraph Customer["Customer boundary"]
    K8S[K8s exporter]
    TF[Terraform / EventBridge ECS]
    GHA[GitHub Action / App]
    SDK[Ingest clients]
  end

  subgraph Evitrus["Evitrus cloud EU"]
    API[API / Ingest]
    CH[(Event store\nEvidence + audit)]
    PG[(Postgres\nIdentity / overlays)]
    UI[Web app]
  end

  subgraph Pull["On demand only"]
    DD[Datadog metrics]
  end

  K8S & TF & GHA & SDK -->|TLS push Evidence Events| API
  API --> CH
  API --> PG
  UI --> API
  API -.->|pull metrics| DD

Data flow (evidence)

Exporters authenticate with a tenant API key and POST events to ingest. The API normalizes each event, seals it with a contentHash, and appends it to the event store under the organisation’s plan retention TTL. Reads derive Changes on demand; the incident view is a client time window over those events. Enrichments and assignments are Postgres overlays, so sealed event bodies are left unchanged.

Network

Public HTTPS is offered for the API, the product app, docs, and the marketing site. Datastores are reachable only inside the cluster and are not exposed on the public internet. Customer exporters initiate outbound HTTPS; the push model does not require inbound firewall holes into the customer VPC.

Optional metric overlays are pulled on demand with customer credentials and are not stored as Evitrus’s system of record.