Observed Signal · Jun 4, 2026 · Technical Release · Source: DEV Community · Impact: 4/5 · Sentiment: Neutral
Google ADK 2.0 Upgrade on Cloud SQL Postgres Issues
A developer report describes practical problems encountered when upgrading Google's Agent Development Kit (google-adk) from 1.x to 2.0 while using Cloud Run with a Cloud SQL (Postgres) session store. Three core issues are detailed: ADK 2.0 uses an async Postgres driver (asyncpg) which requires a 'postgresql+asyncpg://' URL and can break synchronous SQLAlchemy consumers unless the URL is normalized; ADK 2.0 unconditionally reads two new jsonb columns (input_transcription and output_transcription) so missing columns cause session GETs and streaming chat turns to 500; and the legacy v0 (Pickle) schema remains readable but is deprecated and must be migrated to v1 (JSON) into a new database (not in-place) using ADK's migration tooling. The post prescribes a safe deploy order (patch DB first, switch URL, deploy, smoke-test) and provides SQL and command examples.
A Google ADK major-version release changes runtime drivers and database schema behavior; upgrades can silently break chat sessions and require deploy ordering and DB migrations, affecting any deployments using ADK with Postgres.
Track Google Signals & Market Shifts in Real-Time
Polaris7 autonomous intelligence agents track regulatory filings, primary sources, executive changes, and deal flow 24/7. Create your free Explorer workspace to monitor these entities.
Key Takeaways & Evidence Grounding
- google-adk 2.0 switches session service to async and expects an async Postgres driver (asyncpg).
- ADK 2.0 requires DATABASE_URL with scheme 'postgresql+asyncpg://'; synchronous SQLAlchemy engines cannot accept that scheme without normalization.
- ADK 2.0 reads two new events table columns unconditionally: input_transcription (jsonb) and output_transcription (jsonb); missing columns cause runtime 500 errors on session reads.
- The legacy v0 (Pickle) schema remains supported by ADK 2.0 but is deprecated; migrating to v1 (JSON) requires copying data into a new database using 'adk migrate session' (cannot be done in-place).
- Recommended deploy sequence: ALTER TABLE to add new columns first, switch to async URL, deploy ADK 2.0, then smoke-test; migration to v1 can be done later.
Connected Companies & Entities
1 Entity mappedOntology Mapping & Concepts
Related Market Signals & Shifts
Recent verified developments and strategic activity across this market segment.
Google ADK 2.0 Stable Released with Workflow Runtime
Google announced google-adk 2.0.0 as stable, introducing a graph-based Workflow Runtime, an event-first data model, and a structured Task API for agent-to-agent delegation that integrates with the A2A protocol. The release contains breaking changes: 1.x storage and session schemas are incompatible with 2.0 (risking data loss), the agent API and tool decorators are reworked around workflows and Events, and Python 3.11+ is required. The stable build pins LiteLLM to v1.82.6 due to a supply-chain compromise in 1.82.7–1.82.8. The article (published 2026-06-02) includes migration guidance: isolate storage, audit 1.x usage, port agents to the Event model, and update A2A integrations.
Google ADK Master Calls Amazon Bedrock via A2A v1.0
A cross-cloud interoperability benchmark reversed master/worker roles: a Google ADK master on Cloud Run (Gemini 2.5 Flash, us-central1) delegated work to an Amazon Bedrock AgentCore worker (Strands Agents on Nova Micro, us-east-1) over A2A v1.0. The reversal exposed six interoperability defects (five invisible to local tests) across protocol versioning, missing server dependencies, identity federation/IAM trust-policy errors, header forwarding through proxies, and agent-card bind-addresses. On 2026-07-31 (five warm invocations per mode) median latencies were mcp_only 2.96s, a2a_only 18.92s, and verified 16.19s, showing the remote-agent hop costs roughly 6× the local-tool baseline; verified mode runs MCP and A2A concurrently so cost approximates the slower call. The article documents fixes and lessons about testing at deployment seams while noting limited sample size, single region pair, and unresolved policy-scoping questions.
Google ADK Architecture and Core Components
This article presents a high-level architecture and component breakdown for the Google Agent Development Kit (ADK), a framework for building, testing, and deploying AI agent applications. It describes layered architecture including user interfaces, an ADK Runtime & Runner, an agent system (LLM agents, workflows, custom agents), foundational components (session, state, memory, events, tools, artifacts), and a models & knowledge layer that integrates Gemini-family models via Vertex AI or Google AI Studio and supports Retrieval-Augmented Generation (RAG). The piece also covers observability and governance (traceability, logs, metrics, evaluation, alerts, security) and deployment options such as a managed Agent Runtime on Google Cloud, Cloud Run, GKE, and on-premises environments.
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.