16 docs tagged with "platform"

What Delphi records as audit and security events — the per-event fields required by BSR 4.1.1/4.1.2 (time, type, origin IP, user ID), the actor-context block, and which platform actions emit which event.

What init.sh does the first time it runs on a service host — fetch S3 config, resolve env from SSM/SM, decode TLS material, pull images, start containers.

Seed AWS SSM and Secrets Manager from each service's vars.yaml so fetch-env.sh can resolve every variable at boot.

Command-line flags, registry phases, and operator notes for delphi-setup. Run from the Ops host via delphi-setup.sh and the voiceai-db-migrate one-shot container.

Inventory of platform feature flags, what each one turns on or off across services, and how an operator changes them in their deployment.

After migrations and baseline seed succeed, run delphi-setup from the Ops host for first-login setup, upgrades, or optional dev seeds.

Bring up a brand-new Delphi platform deployment from zero — from cloud accounts and config seeding to first call.

How Delphi scrubs PII — DTMF digits, phone numbers, transcripts, and secrets — across three layers: the in-process logger, the DTMF-aware logDTMF helper, and the OTel collector transform processors.

High-level architecture of a Delphi platform deployment — service types, public vs private surfaces, and how data and calls flow between them.

Accounts, DNS, container registry, S3 config bucket, namespace, and tooling required before bringing up a Delphi deployment.

Create one cloud instance per service role with the cloud-init contract Delphi expects, regardless of cloud provider.

Which Delphi platform releases are available, which one to run, and the TelPhi SDK version each release works with.

Export Delphi security and audit events from SigNoz into external SIEM or observability systems.

Concurrent session limits, revocation, and Redis-backed session store behaviour for Delphi Auth.

Platform-managed team secrets and WebRTC TelAPI keys provisioned on team creation — OpenAI and Azure by subscription tier, and what operators must configure first.

Canonical pinned versions of every upstream image and tool the platform ships, and where each version is declared.