Skip to content

1.1.0 - 2026-02-26

Choose a tag to compare

@github-actions github-actions released this 28 Feb 12:53

Release Notes

Highlights

  • Cloud debugging — Observe code running inside Docker containers and remote hosts. No VPN, no port-forwarding. The AI agent connects to a Detrix daemon deployed alongside your service.
  • Multi-expression metrics — Single observation point captures multiple variables simultaneously.

Added

Cloud Debugging

  • Daemon auto-discovery — Clients register with the daemon at startup; the MCP bridge discovers them automatically via advertise_url
  • Virtual File System (VFS) — Transparently fetches source files for the agent. Three configurable sources: bridge (from AI agent workspace), control_plane (from app container), disk (local path). Configurable priority order in [vfs] config section
  • Multi-connection MCP — MCP bridge handles multiple concurrent AI agent sessions with workspace-aware file serving
  • Connection reference counting — Shared DAP connections are safely used by multiple concurrent MCP clients
  • Container restart metric migration — Metrics and connections survive container restarts automatically
  • Admin endpoints — New REST endpoints for daemon administration (/api/v1/admin/*)
  • Docker cloud E2E tests — Comprehensive test suite for Docker-based cloud debugging workflows

Remote App Control

  • wake / sleep MCP tools — Resume or suspend a remote app's debugger on demand. Zero overhead when sleeping; agent wakes the app before adding metrics
  • MCP bridge fallback — If daemon forwarding fails during wake, falls back to direct bridge connection and auto-switches daemon

Authentication & Security

  • Secure-by-default daemon — Authentication enabled by default; set DETRIX_TOKEN env var to configure
  • Per-daemon credential storage — Each daemon instance stores credentials independently; discovery-first auth flow
  • Client ID audit trail — All mutating operations thread a client ID for full audit traceability
  • Hardened client auth — Fixed auth bugs in Go and Rust clients; enabled secure-by-default

Metrics

  • Multi-expression metrics — Single metric can observe multiple expressions simultaneously (expressions: ["symbol", "quantity", "price"])
  • ExpressionValue type with typed projections (numeric, string, boolean)
  • GELF output includes per-expression custom fields (_expr_N_name, _expr_N_value)
  • Configurable max_expressions_per_metric limit (default: 20)
  • Expression merging — Adding a metric at an existing location merges expressions instead of creating a duplicate

MCP Tools

  • disconnect_all — Disconnect all adapters and flush state (new tool, total: 29 tools)

Developer Experience

  • Relative path resolution in observe tool — works with paths relative to project workspace root
  • Smart line suggestionobserve tool suggests the best line when expression is found nearby
  • Build info auto-detection in Go and Rust clients
  • Docker cloud exampleexamples/docker-demo/ shows full cloud debugging setup with Go service, Detrix daemon, and AI agent

Clients (v1.1.1)

  • Python client: detrix-py v1.1.1 on PyPI
  • Go client: github.com/flashus/detrix/clients/go v1.1.1
  • Rust client: detrix-rs v1.1.1 on crates.io
  • All clients support build info auto-detection and secure auth

Breaking Changes

  • WebSocket API: Field names changed from snake_case to camelCase
    (metric_id -> metricId, stack_trace -> stackTrace)
  • gRPC/Proto: expression (singular) replaced with expressions (repeated)
  • REST API: expression field replaced with expressions array
  • Database: Schema consolidated. Requires clean install from v1.0.

Fixed

  • Expression safety validation now enforced on config-file metric imports
  • Go/Rust client auth bugs resolved; secure-by-default now works correctly
  • Inspector no longer returns non-executable locations (function signatures, struct fields) for Go/Rust variable search

Install detrix 1.1.0

Install prebuilt binaries via shell script

curl --proto '=https' --tlsv1.2 -LsSf https://github.com/flashus/detrix/releases/download/v1.1.0/detrix-installer.sh | sh

Install prebuilt binaries via powershell script

powershell -ExecutionPolicy Bypass -c "irm https://github.com/flashus/detrix/releases/download/v1.1.0/detrix-installer.ps1 | iex"

Install prebuilt binaries via Homebrew

brew install flashus/tap/detrix

Download detrix 1.1.0

File Platform Checksum
detrix-aarch64-apple-darwin.tar.xz Apple Silicon macOS checksum
detrix-x86_64-apple-darwin.tar.xz Intel macOS checksum
detrix-x86_64-pc-windows-msvc.zip x64 Windows checksum
detrix-x86_64-pc-windows-msvc.msi x64 Windows checksum
detrix-aarch64-unknown-linux-gnu.tar.xz ARM64 Linux checksum
detrix-x86_64-unknown-linux-gnu.tar.xz x64 Linux checksum