Topaz Weekly Pulse #1: Major Storage milestones and cleaner automation workflows

This week in Topaz: from page blobs to health-checked host orchestration.

This is the first post in a new weekly series: Topaz Ship Log.

Each edition is a concise, case-by-case summary of what changed in Topaz during the week, why it matters, and what it unlocks for local Azure development. This first issue covers the last 7 days of work.

Case 1: CLI and Host are now properly split

Topaz now runs as two separate binaries with a clear contract:

• topaz-host runs the emulator process.
• topaz runs short-lived management commands.

The Host exposes GET /health, and the CLI performs a pre-flight health check before command execution, including validation that both processes use the same working directory. That removes a whole class of confusing "wrong state" failures when scripting across multiple projects.

For automation this is a big quality-of-life improvement: run the host as a service/container, then call CLI commands independently.

Case 2: Azure Storage control plane coverage grew substantially

This week added a large set of Storage account management capabilities, including:

• Storage account management endpoints and commands.
• Listing storage accounts by subscription.
• Name availability checks.
• Account key regeneration.
• Account SAS and Service SAS listing.
• Storage account update (PATCH) support.
• Default endpoint shaping for blob/file/queue/table services.

This closes many of the "day one" gaps for provisioning and managing storage resources with Azure-like workflows.

Case 3: Blob container operations moved much closer to Azure behavior

Container-level behavior was expanded with support for:

• Metadata set/get.
• ACL set/get.
• Lease lifecycle operations (acquire, renew, change, release, break).
• Container properties retrieval.

Combined with related test updates, this improves compatibility for tooling that depends on real ACL and lease semantics instead of simplified placeholders.

Case 4: Blob data plane added key block-blob capabilities

Topaz now supports the core multi-step block-blob flow:

1. Stage blocks (Put Block).
2. Commit block list (Put Block List).
3. Inspect committed/uncommitted blocks (Get Block List).

On top of that, the week included:

• Blob copy operations.
• Blob metadata retrieval.
• Blob properties updates.
• Header correctness fixes (for example Cache-Control handling).
• A broader refactor to use operation result abstractions for more consistent error handling paths.

These changes make SDK and CLI integration paths much more predictable for real-world blob workflows.

Case 5: Page blob support arrived (including page ranges)

Page-blob scenarios were introduced with support for:

• Creating page blobs.
• Uploading pages.
• Reading page blob content.
• Retrieving page ranges.

This is important for workloads and tests that use page-oriented behavior rather than block blobs.

Case 6: Table Storage data plane expanded and protocol alignment improved

Table Storage saw a broad set of updates:

• Table management and entity operations.
• Service properties and stats endpoints.
• CORS preflight handling.
• ACL-related updates.
• Endpoint permission adjustments.
• HTTPS protocol alignment for table endpoints.

The net effect is better fidelity for clients that are strict about protocol and service-level behavior.

Case 7: Auth and request handling hardening

Several cross-cutting correctness improvements landed:

• Authorization checks now include HTTP method in permission evaluation.
• Storage account key generation moved to base64 encoding.
• Router/query parsing and endpoint ordering/path handling fixes for edge cases.

These are not headline features, but they materially reduce subtle interoperability bugs.

Case 8: Faster and more reliable test loops

A lot of work this week was aimed at development speed and CI reliability:

• Terraform test infrastructure improvements (AzureRmBatchFixture, provider pre-initialization).
• Docker-based test setup hardening and host mapping fixes.
• Better diagnostics and output handling in tests.
• Log truncation safeguards to prevent oversized test logs.
• New/updated workflows (including portal build/test/publish paths).

The practical outcome is shorter feedback loops and easier debugging when compatibility regressions appear.

Case 9: Contribution and documentation improvements

The project also gained process and docs upgrades:

• New issue templates and a pull request template.
• Expanded docs for local Terraform and ecosystem integration.
• Clarifications around CLI/Host usage and command expectations.
• Roadmap/backlog updates for upcoming storage and identity work.

These changes make it easier for new contributors to get aligned and for users to understand current feature boundaries.

Case 10: Cross-service API coverage moved forward too

Although Storage dominated the week, several other services also progressed:

• Microsoft Entra ID: tenant-specific token endpoints and group management endpoints were added.
• Key Vault: deleted vault recovery and purge flows were expanded and aligned with expected status semantics.
• Service Bus: namespace model handling improved and network rule set support was introduced.
• Container Registry: repository/tag/blob delete paths, repository listing, replication listing, and HEAD/manifest behavior fixes improved data-plane compatibility.

This matters because most real test environments are multi-service. Improving only one API surface is never enough; the weekly work intentionally moved several surfaces in parallel.

What to expect next

Topaz Ship Log will continue weekly with the same format: major changes, grouped by area, focused on practical impact.

The current direction is clear: deeper Azure Storage parity, tighter CLI/SDK behavior, and continued investment in test-backed compatibility for local-first infrastructure development.