Skip to main content

Thresh 1.6.0: Agent Mode & Hub Connectivity — Connect Your Fleet

· 6 min read
thresh Team
thresh Team
thresh Development Team

Release Date: March 8, 2026
Release Type: Feature Release — Phase 1.6

Thresh 1.6.0 is here, and with it comes the biggest architectural leap since we went cross-platform: agent mode. Every thresh node can now connect to a centralized Thresh Hub over a persistent WebSocket connection, giving you fleet-wide visibility, real-time metrics, and a single pane of glass for all your environments.

🚀 What's New in 1.6.0​

  • 🖧 Agent Mode — Connect any thresh node to Thresh Hub via SignalR WebSocket
  • 🔄 Auto-Reconnect — Agents automatically recover from network interruptions
  • ⚡ Dual Transport — SignalR primary with automatic REST API fallback
  • 🔒 API Key Auth — Secure agent-to-hub authentication
  • 🛡️ TLS Support — Full TLS verification with option to disable for self-signed certs
  • 📊 Real-Time Metrics — Nodes stream CPU, memory, and container metrics to the Hub
  • 🗄️ ConfigurationService — New centralized configuration management for agent settings

🖧 Agent Mode​

The Problem We're Solving​

Running thresh on multiple machines means logging into each one individually, manually checking metrics, and having no unified view of your fleet. Whether you have 2 nodes or 20, this doesn't scale.

Agent mode solves this. Install thresh on any node, point it at your Thresh Hub, and it becomes a managed member of your fleet — instantly visible in the Hub UI with live metrics.

How It Works​

Each agent:

  1. Reads its configuration from ~/.thresh/agent.json
  2. Authenticates with the Hub using its API key
  3. Opens a persistent WebSocket (SignalR) connection
  4. Streams metrics at a configurable interval (default: 30s)
  5. Automatically reconnects if the connection drops

⚡ Getting Started with Agent Mode​

1. Configure Your Agent​

# Set the Hub URL
thresh agent config set midtier-url https://your-hub:7200

# Set your API key (generated in Thresh Hub UI)
thresh agent config set api-key thresh_live_xxxxxxxxxxxx

# For self-signed certs on local hubs
thresh agent config set tls-verify false

2. Start the Agent​

thresh agent start

3. Check Status​

thresh agent status

Example output:

Agent Status
────────────────────────────────────────
Agent ID:    5f6d5891-76d2-466f-a33f-7b87acb17653
Status:      Connected ✓
Hub URL:     https://192.168.4.85:7200
Transport:   SignalR
Uptime:      2h 14m
Last Report: 28 seconds ago

4. View Configuration​

thresh agent config list
MidtierUrl:           https://192.168.4.85:7200
ApiKey:               thresh_live_****
TlsVerify:            false
ReconnectDelay:       5s
MetricsInterval:      30s
AutoFailover:         false
FallbackUrl:          (not set)
Transport:            auto

🔄 Dual Transport: SignalR & REST​

thresh agents use a layered transport strategy:

TransportProtocolWhen Used
SignalRWebSocketDefault — persistent, low-latency
REST APIHTTPFallback if WebSocket unavailable

The Transport setting in agent.json controls this:

{
  "Transport": "auto"     // SignalR first, REST fallback (default)
}
{
  "Transport": "signalr"  // Force SignalR only
}
{
  "Transport": "rest"     // Force REST only
}

High Availability with Failover​

For mission-critical setups, configure a fallback Hub:

thresh agent config set fallback-url https://backup-hub:7200
thresh agent config set auto-failover true

If the primary Hub becomes unreachable, the agent automatically switches to the fallback URL and reconnects. When the primary comes back online, it fails back automatically.

🔒 Security​

API Key Authentication​

All agent-to-hub communication is authenticated with an API key. Keys are:

  • Generated per-hub in the Thresh Hub UI
  • Scoped to a single hub — they don't work cross-hub
  • Stored securely in ~/.thresh/agent.json on each node

TLS​

Production deployments should use a valid TLS certificate on the Hub. For development with self-signed certs:

thresh agent config set tls-verify false
warning

Only disable TLS verification in trusted, private networks. Always use valid certificates in production.

⚙️ Full Agent Configuration Reference​

Configuration lives at ~/.thresh/agent.json:

{
  "AgentId": "5f6d5891-76d2-466f-a33f-7b87acb17653",
  "Enabled": true,
  "MidtierUrl": "https://hub.example.com:7200",
  "ApiKey": "thresh_live_xxxx",
  "FallbackUrl": "",
  "FallbackApiKey": "",
  "Transport": "auto",
  "TlsVerify": true,
  "ReconnectDelay": 5,
  "MetricsInterval": 30,
  "AutoFailover": false,
  "FailoverTimeoutSeconds": 30,
  "FailbackEnabled": true
}
FieldDefaultDescription
AgentIdAuto-generatedUnique node identifier (GUID) — set once, never changes
EnabledtrueEnable/disable agent daemon
MidtierUrl—Primary Thresh Hub URL
ApiKey—Authentication key from Hub UI
FallbackUrl""Backup Hub URL for HA setups
Transport"auto"auto, signalr, or rest
TlsVerifytrueVerify Hub TLS certificate
ReconnectDelay5Seconds between reconnect attempts
MetricsInterval30Seconds between metric reports
AutoFailoverfalseAuto-switch to FallbackUrl on primary outage
FailoverTimeoutSeconds30Seconds before triggering failover
FailbackEnabledtrueReturn to primary when it recovers

🗺️ Coming Soon: Full Hub Integration​

Agent mode in 1.6.0 is the foundation for a much larger vision. Here's what's coming next:

Remote Command Dispatch​

Run commands on any connected node directly from the Hub UI — without SSH. Deploy blueprints, start/stop environments, and run diagnostics remotely.

Fleet Blueprints​

Define blueprints at the Hub level and push them to groups of nodes simultaneously. Perfect for standing up identical dev environments across a team.

Centralized Logs​

Aggregate container and service logs from all nodes into a single searchable stream in the Hub UI.

Role-Based Access Control​

Manage which users can see which nodes, and what operations they can perform, all controlled from the Hub.

Node Groups & Tags​

Organize nodes into groups (e.g., staging, dev, prod) and apply operations to groups rather than individual nodes.

Coming Soon

These features are actively in development. If you're interested in early access, reach out or follow the Roadmap for updates.

📦 Full Release Notes​

New Commands​

thresh agent start                              # Start agent daemon
thresh agent stop                               # Stop agent daemon
thresh agent status                             # Show connection status
thresh agent config set <key> <value>           # Set a config value
thresh agent config get <key>                   # Get a config value
thresh agent config list                        # Show all config

New Files​

  • AgentService.cs — daemon lifecycle, SignalR connection management
  • AgentConfiguration.cs — configuration model and persistence
  • AgentModels.cs — agent status, metrics, and event models
  • ConfigurationService.cs — centralized config service used by agent and MCP

Bug Fixes​

  • Kestrel now binds to 0.0.0.0 on the Hub, fixing agents on separate subnets not being able to connect

⬆️ Upgrading from 1.5.0​

thresh 1.6.0 is fully backward compatible with 1.5.0. All existing commands, blueprints, volumes, and WSL profiles work unchanged.

Download:

# Linux
wget https://github.com/dealer426/thresh/releases/latest/download/thresh-linux-x64.tar.gz
tar -xzf thresh-linux-x64.tar.gz && sudo mv thresh /usr/local/bin/

# macOS
curl -L https://github.com/dealer426/thresh/releases/latest/download/thresh-macos-arm64.tar.gz | tar -xz
sudo mv thresh /usr/local/bin/

# Windows
winget upgrade thresh

Questions or feedback? Open an issue on GitHub or join the discussion.