Changelog

App health#

Real-time visibility into application health is now available directly in the dashboard. When something goes wrong with your app - whether it's an out-of-memory crash, a segmentation fault, or a misconfigured start command - you'll see an alert banner at the top of the page with actionable details.

Click the banner to open the App health panel, which shows:

• Instance status at a glance - See the state of every running instance across all your process types (web, worker, cron, job), including restart counts and exit codes.
• Automatic issue detection - Recognized failure scenarios include OOM kills (exit code 137), segmentation faults (exit code 139), permission errors, missing binaries, application errors, and health check misconfigurations.
• Troubleshooting guidance - Each detected issue comes with a plain-language explanation and specific steps to resolve it, so you can diagnose problems without digging through logs.
• Live updates - Instance statuses refresh in real time, so you can watch recovery happen as you apply fixes.

The alert banner appears across all application pages, so you'll never miss an issue regardless of where you are in the dashboard.

Usage page#

The new Usage page gives you full visibility into your resource consumption and costs directly in the dashboard.

Get a clear breakdown of spending across all your services - applications, databases, static sites, object storage, and load balancers - with daily cost tracking and period-over-period comparison.

Key highlights:

• Summary cards - See your current period total, projected end-of-period cost, and previous period comparison with percentage change indicators at a glance.
• Interactive charts - Visualize daily or cumulative costs with stacked bar and area charts, color-coded by usage type.
• Flexible filtering - Filter by project, service, or usage type to drill into exactly the costs you care about.
• Detailed cost table - Expandable table showing per-service net, gross, and tax amounts grouped by usage category, with deleted services clearly marked.
• Export and print - Download your usage data as CSV or JSON, or print a full report for your records.

The Usage page is accessible from the Settings section in the sidebar.

Sevalla Terraform Provider#

The official Sevalla Terraform Provider is now available. Define and manage your entire Sevalla infrastructure as code using HashiCorp Terraform.

Provision applications, databases, static sites, and configure domains, environment variables, and scaling - all through declarative HCL configuration files. Perfect for teams that want reproducible, version-controlled infrastructure.

Sevalla CLI#

The official Sevalla CLI is here - manage your entire Sevalla infrastructure from the terminal.

Deploy apps, provision databases, manage static sites, configure domains, and monitor your services without leaving the command line. Every action available in the dashboard is now at your fingertips via a clean, consistent command structure.

Install it with your preferred package manager:

# Homebrew
brew install sevalla-hosting/tap/sevalla

# npm (Node.js 22+)
npm install -g @sevalla/cli

# Shell script
bash <(curl -fsSL https://raw.githubusercontent.com/sevalla-hosting/cli/main/install.sh)

Key highlights:

• Full platform coverage - Applications, databases, static sites, domains, and monitoring all in one tool.
• Scriptable output - Every command supports --json for easy integration with CI/CD pipelines and automation scripts.
• Flexible authentication - Interactive browser-based login for local use, or API token via SEVALLA_API_TOKEN for automated workflows.

Run sevalla --help to explore all available commands, or check out the GitHub repository for full documentation.

Sevalla MCP server#

The official Sevalla MCP (Model Context Protocol) server is now available. You can now give AI agents and coding assistants full access to the Sevalla PaaS API.

You can seamlessly connect popular MCP clients like Claude Code, OpenCode, Cursor, and other coding agents directly to our hosted server at https://mcp.sevalla.com/mcp. Authentication is handled effortlessly via OAuth, meaning you don't need to hardcode API keys in your configuration files.

Now that your AI can spin up and manage infrastructure on command, the only limit is your imagination. Well, that and your prompt engineering skills. 😉