roadmap: IISA v1 completion plan #528

Overview

This issue documents the execution plan for completing IISA v1 and related improvements. The remaining work is organized into 5 sequential waves, with each wave building on the previous.

Wave Structure

Wave 1: IISA HTTP Hardening (7 issues)

Establish production-ready HTTP communication before deployment.

• feat(iisa): Add timeout and retry logic to HTTP client #507
• feat(iisa): Define error handling and resilience policy #514
• feat(config): Expand IISA configuration options #519
• test(iisa): Add comprehensive HTTP client tests #509
• security: Review and harden IISA HTTP communication #517
• feat(observability): Add logging, metrics, and tracing for IISA HTTP client #518
• docs: Update documentation for containerized IISA architecture #508

Rationale: The HTTP client is the critical integration point. Timeout/retry (#507) and error handling (#514) are foundational. Security (#517) and observability (#518) follow. Configuration (#519) exposes these settings. Tests (#509) validate the implementation. Documentation (#508) captures the final state.

Wave 2: Deployment & Release (2 issues)

Ship the hardened IISA to production.

• release: Plan and execute IISA containerization release #516
• infra: Add Kubernetes deployment configuration for dipper and IISA #515

Note: These issues are tracked here for coordination, but the implementation work happens in edgeandnode/subgraph-dips-indexer-selection. See also subgraph-dips-indexer-selection#1 for related production hardening tasks in that repo.

Rationale: With the HTTP client hardened and tested, the system is ready for deployment. K8s config (#515) enables deployment, and the release (#516) coordinates the rollout.

Wave 3: Agreement Lifecycle (2 issues)

Improve reliability and performance of agreement management.

• perf: Add database-side aggregation for pending agreements query #526
• Add periodic reassignment loop for active indexing requests #524

Rationale: Post-deployment improvements. The reassignment loop (#524) handles indexer failures gracefully. DB aggregation (#526) optimizes the batch query introduced in PR #522.

#524 replaced with #550

Wave 4: Scalability Validation (1 issue)

Validate the system under production-like load.

• test(iisa): Validate HTTP payload scalability at 25K subgraphs / 100 indexers #520

Rationale: With the system deployed and stable, validate that it handles the expected scale. This informs any further optimization needs.

Wave 5: Future Work (1 issue)

Deferred improvements for v2.

• Stale Created agreements lack timeout mechanism #525

Rationale: This is a v2 feature that addresses edge cases in agreement lifecycle. Lower priority than core v1 functionality.

Completed

• Optimize CI pipeline performance #523 CI pipeline optimization (PR perf(ci): parallelize clippy and adopt cargo-nextest #527)
• feat(iisa): add load balancing context with batch query optimization #522 Extend CandidateSelection trait for load balancing context

Dependencies

Wave 1 ──► Wave 2 ──► Wave 3 ──► Wave 4 ──► Wave 5
              │
              └── Work in subgraph-dips-indexer-selection repo