Changelog

• AI Engineer is Coming to Paris
• New Tools for Koyeb MCP Server
• Koyeb CLI: Add '--wait' Flag
• Use Dia 1.6B to Build a Text-to-Speech Application on Serverless GPUs

• Koyeb MCP Server in Beta
• Flux.1 Kontext [dev] one-click available
• New redesigned and personalized onboarding flow
• New tutorial: Use PydanticAI to Build AI Agents

• Public TCP Support Private Preview
• Koyeb Partner Hub is Live
• Logs now use UTC
• Join us at RAISE Summit and AI Developer Meetup in Paris

• 60% Faster Time to Healthy and Deployment Promotion
• Improved Traffic Shifting During Deployment Promotion
• Serverless Postgres Now Included in Usage Summary
• Magistral, DeepSeek-R1 0528 Qwen3 8B, Llama 4 Scout Available in One-Click

• New Go, Java, Node.js, Python, PHP, and Ruby Versions
• Multi-Agent Hackathon: Build Multi-Modal Workflows with Tenstorrent
• ResembleAI Chatterbox Available in One-Click

• Newly Created Services Now Accept Traffic Immediately
• Fixed Instance Healthiness Reporting During New Deployments
• Mistral Devstral Small Available in One-Click

• Serverless Postgres General Availability
• Better Visibility of Replica Health and History
• New One-Click Models: Qwen 3 235B, 32B, 14B, and 8B
• Filter Organizations by Name in the Organization Selector

• Improved Visibility of GPU Availability
• Improved Billing Alert Configuration
• Enhanced Organization Selector
• Fixed Environment Variable Interpolation

• Tenstorrent Private Preview: New One-Click Models & Fix for Init/Reset Delays
• Improved Errors on Deployment Failures
• Control Panel: Improvements & Fixes
• Meet the team at AI Rabbit Hole in the Bay Area next week!

• NVIDIA L40S now in North America
• Billing Alerts
• Tenstorrent Private Preview: Updated tt-smi and Tenstorrent Kernel Modules
• AI Rabbit Hole in SF: Come Meet the Team

• Control Panel: Real-Time GPU Availability
• Control Panel: Self-Serve Coupon Redemption
• LLM x Law Hackathon at Stanford University & GOSIM AI Conference in Paris

• Increased Maximum Grace Period Limit for Health Checks
• Control Panel: Fixed Display Bug During Wake and Scale-up
• Control Panel: New Sign-up and Sign-in Pages

• A100 SXM available in North America
• New One-Click Models: Gemma 3, Phi-4 Multimodal, and Mistral Small 3.1
• Custom Scale-to-Zero Idle Timeout for Pro Plans and Above
• Best Open Source Multimodal Vision Models in 2025

• Partnering with Vultr For Serverless & Global AI Deployments
• Scale-to-Zero Enabled on Free Instances
• Control Panel: Access log history
• Control Panel: Automatic Region Selection

• New One-Click AI Models: QwQ 32B, R1 1776 Distill Llama 70B, and more
• 2x, 4x, 8x H100 GPU Instances available on request
• Control panel: New instance and regions selector

• Tenstorrent Cloud Instances in Private Preview
• Control Panel: Enhanced Docker Command & Entrypoint Overrides
• Multi-region Deployment: Improved Traffic Shifting
• Control Panel: Improve Service Redeploy

• Nvidia A100 and H100 GPUs Available in Dallas
• Custom Request Timeout for Scale Plan Users
• Deploy Larger Docker Images to GPUs
• Next-Gen AI Infra with Tenstorrent & LlamaIndex in San Francisco

• Qwen 2.5 VL 7B Instruct and Qwen 2.5 VL 72B Instruct One-Click Models
• Improved Scale-to-Zero Cold Start
• Best Open Source LLMs in 2025

• Nvidia A100 GPUs Available in US
• Improvement: Wake Up During New Deployments
• AI Agents Hackathon with AI Tinkerers & Anthropic in Paris

• DeepSeek-R1 Qwen 32B, Llama 70B and Mistral Small 3 One-Click Models
• DeepSeek-R1’s Multi-Lingual and Agentic RAG Capabilities in Practice
• Control panel: Bulk delete domains

• New AI models in one-click app catalog
• Faster out-of-memory detection
• Fixed autoscaling issue

• New plans: Pro and Scale
• Deploy AI models in one click
• Control panel: Fixed issue with CPU usage
• Control panel: Bulk delete secrets

• Scale-to-Zero in Public Preview
• A100, L40s, and L4 price drops
• Enhanced experience with service mesh
• Enhanced experience with Volumes created from Snapshots

• Docker image download progress
• New website
• Control panel: Discard pending service changes
• Control panel: Easily duplicate Services

• Volume Snapshots in Public Preview
• Enhanced experience in service scaling
• Bulk import secret
• New tutorial: Use FLUX, PyTorch, and Streamlit to Build an AI Image Generation App

• Faster Dockerfile-based builds
• New Go, Node.js, PHP, Python, and Ruby versions
• Improved experience when staging service updates
• Enhanced experience for Koyeb GitHub app installation

• Multi A100, RTX A6000, and L40S GPUs available on request
• CLI Update: Show service changes staged for next deployment
• Improved PORT experience
• New tutorial: Use Stable Diffusion and PyTorch to Build an Image Inpainting Service

• Performance improvements for Dockerfile-based builds
• Fixed secret interpolation at build time
• Nvidia A100 GPUs in self-service for Starter organizations
• New tutorial: Fine-Tune MistralAI and Evaluate the Fine-Tuned Model on Koyeb Serverless GPUs

• Improved build status
• Easier environment variable creation
• Fixed KOYEB_SERVICE_PRIVATE_DOMAIN variable value
• New tutorial: Fine-Tune Llama 3.1 8B using QLORA on Koyeb Serverless GPUs

• Volumes: Fixed permission for non-root users
• Update the settings of paused Services
• Autoscaling: Decrease the maximum scale at any time
• New tutorial: Using ComfyUI, ComfyUI Manager, and Flux to Generate High-Quality Images on Koyeb

• Faster deployments for new Hobby users
• Control panel: Improved deployment navigation experience
• Control panel: Access deployment definition from deployment overview
• Control panel: Fixed issue for logs on deployment without rebuild

• Paris Upgrade: 70% Faster Deployments & More
• Select the Koyeb App on database Service creation
• Deploy Open WebUI, Shiny, and Argilla in one click

• Multi-line secrets in the control panel
• Clear build cache on redeploy
• Fixed connection issue with koyeb instance exec and cp
• Deploy Tailscale SSH, Label Studio, and PGMQ in one click

• CLI 4.3: Improved experience with logs, grace period on health check, and more
• Log exporter 1.2
• New one-click app: Deploy PyTorch Jupyter Notebooks

• Save service configurations
• New organization’s activity in the control panel
• New tutorial: Using YOLO for Real-Time Object Detection with Koyeb GPUs

• Volumes in the control panel
• Increased Docker image size for GPUs
• Skip build on redeploy
• New tutorial: Using OpenAI Whisper to transcribe podcasts on Koyeb GPUs

• Control panel: New Service overview and deployment navigation experience
• Control panel: Improved troubleshooting for a failed deployment with contextual recommendations
• Control panel: Enabled updates for services deployed from a local directory
• New tutorial: Deploy Redis as an in-memory database

• Autoscaling GA: Scale Fast, Sleep Well, Don't Break the Bank
• GPUs Public Preview: Run AI workloads on H100, A100, L40S, and more
• Volumes Technical Preview: Blazing-fast NVMe SSD for Your Data
• AWS Regions: Build, Run, Scale on AWS with Koyeb
• Koyeb for Startups: Accelerate with Credits for High-Performance Infrastructure

• Serverless RTX GPUs in technical preview
• Autoscaling: Scale based on concurrent requests
• Show secret values in the control panel
• New tutorial: Deploy Infisical on Koyeb

• Exclude files to trigger Deployments with .koyebignore
• Add Instance replica index
• How to Deploy Strapi as a Content Backend and Headless CMS on Koyeb

• Deploy from a local directory
• Reference Koyeb variables in your environment variables
• Deploy specific Git commit via the CLI
• Build a Video Processing Pipeline with AssemblyAI on Koyeb

• New Control Panel in preview
• Rename Apps and Services in the new control panel
• Using LlamaIndex and MongoDB to Build a Job Search Assistant

• Apps and Services rename
• Edit App public domain
• New CLI version
• Use Autonomous AI Agents with SerpApi and AutoGPT to Build a Trip Planner

• Serverless GPUs in Private Preview
• Performance improvements on API endpoints
• New one-click apps and tutorials: Directus, Flowise, and Wasp

• Add platform detection for Docker images
• Improved error message in case of an invalid Docker command
• CLI: Improved service configuration management, access metrics, and more
• Using Groq to Build a Real-Time Language Translation App

• Introducing the command palette for faster deployment
• Deploy without rebuild in the control panel
• Improved scale-up during traffic spikes
• Use LlamaIndex to Build a Retrieval-Augmented Generation (RAG) Application

• New service form available for everyone
• Faster builds
• Koyeb environment variables at build time
• Using Astro and Server-Sent Events (SSE) to Build Realtime In-App Notifications

• Skip the build on new deployments
• Improved promotion of new deployments
• Improved database service usage reporting
• Build a Multimodal Chat App using LLava, Chainlit, and Replicate

• New deployment events
• Improved deployment time after build succeeded
• CLI: Logs include timestamps
• Deploy Hapi and Beego in one click

• Faster build and runtime logs
• Improved initial deployment lifecycle
• Fixed public data transfer metric
• Build a Retrieval-Augmented Generation Chatbot using pgvector

• Autoscaling is in Public Preview
• Singapore Region is now Generally Available
• Empty Secrets are allowed in the control panel
• Deploy LlamaIndex in one click

• Improved Docker-based deployments
• New service configuration form rollout update
• CLI: Docker image validation, display Git SHA, and more
• Deploy LangServe by LangChain in 1-Click

• Service Metrics Instances History
• A Git commit SHA for all git-based deployments
• Deploy to Koyeb button improvements
• Use MistralAI, FastAPI, and FastUI to Build a Conversational AI Chatbot

• Autoscaling operations available in activities
• New service configuration form
• Autoscaling support in Terraform provider
• Deploy Ollama in one click and HTMX + PostgREST tutorial

• CLI: Database instance type, and more
• New Service Mesh: 50% faster and Rollout Update
• No more delay on authentications and organization switch
• Deploy Hono and Flipt in one click

• Eco instances available in Singapore
• CLI: Autoscaling, skip cache
• New Pricing Page including Eco, Standard, and Postgres instances
• New DeepSparse, Express, and Spring Boot in one click

• Autoscaling in the control panel - Technical Preview Open
• Larger Postgres database instances available
• New dashboard available for everyone
• New gRPC and Spring Authorization tutorials

• Autoscaling available in technical preview
• New Go, Node.js, PHP, Python, Java, and Ruby versions
• Support for Git submodules

• Serverless Postgres Pricing with up to 4GB of RAM per database
• CLI updates: Copy files to and from Koyeb Instances, container registry credential
• Serverless Postgres roles and logical databases sync
• Metrics selector improvement
• New deploy docs and tutorials

• New Eco Instances
• Enhanced Free Instance
• Billing alerts
• Baselime Log Exporter One-Click App

• Koyeb-provided runtime variables
• Rollout of our new service mesh for faster deployments
• New control panel home page preview
• Use pgvector + Hugging Face on Koyeb

• Serverless Postgres Public Preview
• Estimated cost based on Service configuration
• New 1-click apps and content to help you get started
• Added Postgres Database Service support in the CLI

• 1-Click Apps Catalog
• Per region latencies displayed in Service creation
• Managed Postgres Database - Connect to your applications
• Deploy apps using Docker Compose

• Managed Postgres Database - Role Management
• Terraform provider enhancements
• Launch Shell Commands with the Right Execution Environment
• Deploy a NestJS App
• Deploy Preview Environments on Koyeb for GitHub Pull Requests

• We raised $7M to simplify serverless app deployments globally
• Managed Postgres Database- New regions
• Automatic PORT Variable Injection at Runtime
• New Free Tier

• Managed Postgres Database - Logical database management
• Run your container in privileged mode
• Learn how we built the global deployment engine

• Managed Postgres is available in private preview
• New Free instance type

• San Francisco, Paris, Tokyo, and Singapore regions are in public preview
• Managed Postgres private early access is starting
• Return of the free plan
• New guide to deploy Uptime Kuma on Koyeb
• Use cache to accelerate Docker builds blog post

• Koyeb CLI 3.2.0
• Improved region management in the CLI
• Full dotenv support for bulk environment variable import
• Deploy Bun on Koyeb
• Deploy Cal.com on Koyeb

• Koyeb Metrics is in GA
• Improved region management in the CLI
• Fixed logs issue in the CLI
• New guide to deploy DocuSeal on Koyeb
• Use cache to accelerate Docker builds blog post

• Dockerfile deployments on high-performance microVMs is in GA
• Koyeb Organizations in the CLI
• New guide to deploy Authorizer on Koyeb
• Fixed dipslay bug on new git deployment
• Koyeb at Infobip Shift’s Conference and on the latest Cup o’ Go podcast

• Workers in GA: deploy and scale high-performance background jobs
• Zero downtime redeploys: seamless networking during redeploys
• Billing widget to monitor usage in the control panel
• API credentials for Organizations
• Improved logs experience

• New Node.js and Maven versions
• Cancel deployments in progress from the deployments list
• Organization deactivation
• Deploy Grafana tutorial

• New San Francisco, Paris, Tokyo, and Singapore regions in early access
• Koyeb CLI 3.0
• .env support in control-panel: bulk add environment variables

• Build and runtime log exporter documentation
• New Polyscale x Koyeb documentation
• Improved the console in the control panel
• New tutorial: Build and deploy a gRPC-Web App using Rust Tonic and React

• Improved deployment information in the control panel
• New Remix, FastAPI, Express, and Django apps in the control panel
• Koyeb Build and Runtime Log Exporter in technical preview
• Download usage CSV for a specific period
• Node.js version 20.4.0

• New deployment experience in the control panel
• New documentation plaform and structure
• Minor control panel improvements for building from a Dockerfile and managing custom domains

• Node.js 20.3.1, 18.16.1, and 16.20.1
• Prometheus exporter for Koyeb resources by Daz
• New doc for running Backstage.io on Koyeb

• New service experience in the control panel - Private preview
• Cancel ongoing deployments improvement
• Koyeb at MongoDB.local NYC and on Cup o’ Go!

• Execute commands on your service instances using the CLI
• Published What is a MicroVM? blog post
• Improved custom domains documentation

• Build from Dockerfile support in the control panel
• Improved deployment experience
• We added support for Node.js 20.2.0 and Yarn 4.00-rc.44
• New Upstash Redis database integration documentation

• Build from Dockerfile is in technical preview
• Monthly usage summary
• New Remix deployment guide

• We added support for the latest version of Node.js 20.1, Yarn 3.5.1, and 4.0.0-rc-43
• Improved custom domain status and messages
• New Next.js and Nuxt deployment guides
• Google container registry and secrets improvements

• 99% faster network propagation
• Koyeb Pricing goes live in June
• Improved automated mail on deployment failures

• 35% faster deployment time
• pnpm, Node.js 20, and Java 20 support
• New Django deployment guide and What is gRPC? blog post

• Improved build error message on app detection failure
• New resources: Fastify deployment guide, Socket.io tutorial for using WebSockets, and an eBPF blog post
• Koyeb Terraform Provider improvements and fixes
• Control panel improvements

• Added end-to-end testing to our example applications
• API access token improvement
• Login error message improvement
• Collecting feedback

• First deployment experience improvement
• Deployment status improvement
• Custom domain improvement
• Group instances by deployments

• gRPC and HTTP/2 support in the control panel
• Cache invalidation on redeployments
• Python Celery tutorial for deploying background workers

• Organizations and Multi-user features in public preview
• Deploy from a sub-directory in a git repository
• PlanetScale database documentation

Changelog update #8:

• Display new 503 page to indicate the deployment is in progress
• Published new health checks documentation
• The whole team came together during our bi-annual Koyeb Offsite last week

Changelog update #7:

• Custom health checks in public preview
• Deployment status improvements
• New FastAPI deployment guide

Changelog update #6:

• Native worker support - one of the most demanded features on the platform - is now in public preview
• You can now create multiple Organizations (and Teams!)
• Community members reported unexpected timeouts on active WebSocket connections. We released a patch to fix this issue. WebSocket connections can now be active for several hours
• We updated our Node.js Buildback to support the latest versions of Node.js, including Node.js 19

Changelog update #5:

• We added gRPC and HTTP/2 support to our edge infrastructure, you can now query your public services over the Internet using gRPC and HTTP/2
• New Aiven x Koyeb database documentation to help you get started using Aiven MySQL databases with your apps running on Koyeb
• Meet us at MWC Barcelona this week

Changelog update #4:

• Koyeb Organizations is in technical preview. Invite your teammates to collaboratively work on all your apps and deployments
• New Neon x Koyeb database documentation
• Meet us at MWC Barcelona next week! Come say hi at our booth and collect free swag

Changelog update #3:

• Public release of the Koyeb Terraform provider
• New health checks status messages
• New MongoDB Atlas x Koyeb documentation
• Fix logs when switching between new deployments

Changelog update #2:

• User-defined health checks are in technical preview. We support HTTP and TCP health checks. Reach out to us if you want to get early access.
• Apps Status and Apps listing revamp in the control panel
• New CockroachDB x Koyeb documentation
• Fixed a UI bug for deployments from a private registry

We hope your first month of 2023 was eventful and serverless! 😉

Kicking off the first edition of this new and improved changelog, allow us to introduce:

• New build and runtime logs components
• New way to manage your custom domains
• The ability to pause (and resume!) Apps

Today, we are excited to announce the Koyeb Pulumi provider! Pulumi is an open source infrastructure as code tool for managing your infrastructure. You can now use Pulumi and your favorite programming language to provision and Koyeb resources including Apps, Services, Domains, and Secrets.

Run closer to your users in the US with our new core location in Washington, DC! You can now deploy your applications in the world's most interconnected region with one click, CLI command, or API call.

You can now deploy your full-stack apps and APIs in the heart of Europe with our Frankfurt core location. Your application will be up and running in microVMs on our high-performance BareMetal servers with one click, CLI command, or API call.

We added the ability to enable and disable automatic deployments of your application every time you push a change. Previously, each change you pushed created a new deployment.

We improved the creation flow for Apps and Services, which is now composed of three simple steps: select your deployment method, import your project, and configure your service. We also added helper links at each step so you can easily learn more about Koyeb concepts during the service configuration process.

Last month, we added example applications to the control panel to give users a faster way to see Koyeb in action. Previously, they were only visible for users without any active apps. They can now be accessed at anytime via the control panel.

We published documentation to explain how to connect a managed Postres database from Crunchy Bridge to an application running on Koyeb. Documentation covering more database providers is coming soon.

You can now request the regions where you want to deploy your apps directly from the service creation page in the control panel. Some of these are coming very soon to the platform. Any ideas which ones?

When listing your repositories, we now display the repository’s visibility and last change date. This makes it easier to find the repositories you most recently worked on.

We added the ability to cancel an ongoing deployment. This is helpful when you have deployments queued behind the ongoing deployment and you do not need it anymore.

There is now a maximum limit for concurrent provisioning deployments per service to one. This limit can be increased for Startup plan subscribers by contacting us.

Our free tier includes a $5.5 per month credit that can be used for Nano and Micro instances. Previously, we displayed the prices of these instances, causing confusion for those using the free tier.

We started adding example applications to the console, so you can discover the Koyeb experience sooner. These are minimalistic apps with more advanced examples to follow.

Previously, deleting a Koyeb App required you to manually delete all associated Services first. Now, you can delete an App and all its Services directly from the App's setting page or using the CLI.

We improved the repository and branch selectors by adding visual feedback to indicate when we are retrieving your GitHub repositories data. Previously, no visual indicators were provided causing confusion.

You can now automatically manage your account subscription to the Starter or Startup plan right from the Koyeb control panel.

We published documentation to go alongside the public release of our metrics feature. We dive into the explations about the different metrics you can view in the console.

We are happy to share that we now display several metrics for services running on Koyeb. You can now view the CPU usage, memory usage, responses time, requests throughput and pulic data transfer under the metrics tab of your service.

We added a new Golang deployment guide to our documentation. This will help users learn how to deploy their Golang apps with using the console or the CLI.

We shared a new Rails deployment guide to our documentation. This will help users learn how to deploy a Rails application on Koyeb using either the CLI or the console.

We published a new deployment guide for Python Flask applications. This will help users learn how to deploy a Flask application on Koyeb using either the CLI or the console.

We now natively support Clojure applications. Simply connect your GitHub repository and Koyeb will automatically build and deploy your application.

Some users experienced build failures when deploying a Ruby app on the platform. This issue is now mitigated.

Previously, the Golang version defined in the go.mod file was ignored resulting in inconsistency during the build and deployment of Go applications.

We created a guide and video to help users migrate their backend and full-stack applications from Heroku to Koyeb.

We publicly released the Koyeb Terraform provider. The provider allows you to provision, change, and version Koyeb resources including apps, domains, secrets, and services.

Each time you push a new commit to your GitHub repository, you can now see and follow the Koyeb deployment status for each GitHub commit.

We improved the service and deployment documentation to make it easier for you to get started, learn how Services work, and troubleshoot issues.

Setting up build and run commands for git-driven deployments was previously only available using the Koyeb console. You can now pass them when using the Koyeb CLI.

You can now pause a service while keeping the configuration and be able to restart that service later on. This is especially useful during the development and testing phases.

When deploying your apps with git, we now show the Koyeb deployments status on your GitHub repository's page.

Previously, the branch had to be set via a free-text input when using git-driven deployment. This action was error-prone. We now list all the branches based on the selected repository.

When using git-driven deployment, the branch is now automatically set to the default repository branch.

All versions of the CLI are now also shipped for ARM64 machines and in brew on MacOS.

We've added automatic retry on builds to make sure hardware failures are transparent from an end-user perspective.

You can now easily switch the Koyeb App your domains are pointing to.

All incoming traffic will be rerouted to the new Koyeb App immediately.

Inside of the service mesh, our DNS servers were returning NXDOMAIN status on IPv6 DNS resolution causing resolution failures with some DNS clients. They now return NOERROR and the resolution performs as it should.

We now allow the use of APEX domains (top-level domains) with your Koyeb Apps.

To bind APEX domains on Koyeb, your DNS provider needs to support the creation of ALIAS/ANAME records.

The time limit for builds is now set to 30 minutes to provide faster feedback and prevent builds from hanging in case of build misconfiguration.

If your build needs more than 30 minutes to complete, contact us!

The Koyeb engine and APIs received a major update to prepare for the addition of new core locations. You can now configure instance types, scale, and environment variables at the service level and per region.

A new version of the CLI and the Koyeb console have been released to reflect these changes and provide a simple way to deploy across multiple regions.

You can now set up your own domains on your Koyeb Apps and enjoy fully automated TLS. The feature is currently available in preview, contact us if you want to set up your domain!

We’re excited to share that the new Koyeb Edge Network is live and WebSockets support is available for everyone without any modification. WebSockets can now be used on all apps.

The changes we made on our Edge Network also bring the number of edge locations available on the Koyeb platform to over 270 cities (55 previously) to improve latencies and accelerate load times around the world.

We added search to the Koyeb documentation to let you easily find the information you are looking for.

We revamped the Koyeb Edge Network to support Websockets, custom domains, and additional core locations.

We identified a bug where old unhealthy deployments were still running instead of being stopped. Unhealthy deployments are now stopped, freeing up resources for other deployments.

Koyeb will now send you an email to inform you in case of a deployment failure.

You can now execute commands or get a shell to troubleshoot your deployments via the control panel. Previously this feature was only accessible using Koyeb CLI.

We completely revamped the Services status to reflect the reality. Our new status system is now available for Instances, Deployments, and Services.

We have enriched the activity stream by adding operations performed on Apps and Services events so you now get the full history of actions performed on your account

We accelerated the build for applications deployed using git-driven deployment by 10-20 seconds. The service mesh is no longer connected during the build process, which reduces overhead.

We have added new instance types to support intensive workloads. You can now run your applications on the Large (4 CPUs / 4GB RAM) and XLarge (8CPUs / 8GB RAM) instances. You can request access to these instances directly from the Koyeb control panel.

We completely revamped deployments status. You can now know the exact state of your deployments from a single place, so you don’t need to look at different levels to troubleshoot and identify errors. Each deployment has a status and a clear message defining its current state.

We enhanced the behavior of Node.js builds with devDependencies. Previously, if you set NODE_ENV to production, devDependencies were not installed, resulting in errors for some applications. We now install all project dependencies during the build stage.

We added CLI helpers throughout the Koyeb Console to show how to use the Koyeb CLI to interact with the console directly from your terminal.

The Deploy to Koyeb button helps users deploy applications with one single click. It's ideal for developers looking to create fast deployment option for their users. This button can easily be added to Docker registry public descriptions and git repositories readme files.

For advanced troubleshooting cases, you can now execute commands on instances supporting your services. You can execute simple commands or launch an interactive shell to investigate production issues.

Build and runtime logs were previously only accessible via the Koyeb control panel. We added them to the Koyeb CLI: you can now visualize your logs directly from your terminal.

The Koyeb CLI now supports the creation of Services using git-driven deployment. The GitHub app must be first installed on your account using the control panel to use this feature.

We released major improvements of the API and a new version of the Koyeb CLI to improve the developer experience. It's now easier to enjoy the full Koyeb experience right from your terminal!

Services are backed by one or more Instances which are technically Firecracker MicroVMs. These Instances are automatically created by the platform to support your application. You can now visualize the Instances history for each Service and for each Deployment.

The Instance history is useful to troubleshoot deployment failures: you can easily retrieve runtime logs and execute shell commands on a specific Instance.

We improved the build and runtime logs control panel components with a fullscreen mode and improved performances for applications with a high logging rate.

When using Koyeb's built-in Continuous Deployment feature, we now display the build duration to help you analyze and optimize your build performances.

Previously, there were no clear start or end messages in the build logs.

We now display clear messages to communicate when a build starts, succeeds, or fails. This information is useful when troubleshooting build issues.

We see security as a crucial subject and we believe this should be a built-in feature, not an option.

To prevent data leaks, we already encrypt all in-transit data, including the network traffic between all of our hypervisors. We’re acting as if we had absolutely no trust in the underlying private networks we use.

We now also encrypt all customer data stored at rest, without any action required from you. All customer data stored on our hypervisors are now encrypted using AES256.

We improved our Deploy with Git documentation by adding sections explaining how to use the build and run commands on Koyeb.

We created a collection of application examples you can deploy on Koyeb with a single click. These examples allow you to discover how to deploy applications using different runtimes and frameworks.

Previously, it was impossible to access the runtime logs for an entire Service because you had to switch between nodes, stdout, and stderr streams to see all the logs, making it hard to troubleshoot issues.

You can now see all runtime logs for a Service in a single view and filter the logs by node.

We improved the service overview in the console to simplify deployment troubleshooting. Only active and in-transition deployments are now visible and the previous deployments can be found in the dedicated tab.

We also added the deployment ID near the node entry in the node table as you might have nodes related to different deployments running simultaneously when traffic is shifting.

Secrets for container registry configuration were previously simple Key/Value objects. Now, there is a typed secret available directly in the API and CLI that simplifies configuring your private container registries.

This allows us to validate the structure of your registry secrets before you submit them, only list registry secrets when you are adding a private registry and list only simple secrets when you are configuring environment variables.

We now display the name of the API token which performed each action listed in your recent activities.

For those who like working in the dark, we added a dark mode to the console.

We revamped our console to give it a more sleek look and feel.

We added Trace IDs to the Koyeb 5XX error pages that you might encounter if your application can't be reached from the Koyeb network. This unique Trace ID will help us diagnose and troubleshoot issues faster.

For git-driven deployments, it was previously not possible to run custom commands during the build stage and having a Procfile in a repository was mandatory to let Koyeb know how to run your application.

Git-driven deployment is available in preview! You can now connect your GitHub repository in a matter of seconds, and Koyeb will automatically build and deploy your application each time you git push.

We now pre-set a PORT environment variable matching the default service port.

The log control panel now renders ASCII colors to improve readability and simplify troubleshooting. This feature request was tracked here.

Deploying from a DigitalOcean registry previously required you to manually configure a Secret for your credentials.

Koyeb CLI version 2.0.1 has been released to improve the developer experience when interacting with Koyeb from the terminal.

When using the service discovery via the mesh, you previously had to copy the name of each service manually. This action was error-prone.

We've observed an increased amount of failures related to the growing number of deployments.

We reworked several core deployment jobs to increase deployment speed and reduce unexpected failures.

Previously, retrieving the nodes' status details in the control panel was done by hovering over the status. It was both counter-intuitive and unergonomic.

You now view the status and easily understand what is happening on each node thanks to a dedicated panel.

To deploy from an Amazon Elastic Container Registry, you previously had to manually create a new Secret to store the AWS ECR configuration. This was error-prone and often led to deployment failures.

We simplified the process by adding AWS ECR to the list of registry providers on the service creation page allowing you to easily deploy Docker images from AWS ECR.

The GitHub Action for Koyeb CLI offers a simple way to install, configure and use the CLI in your GitHub Actions workflows.

The Action is useful to add continuous deployment of your app to the Koyeb platform. It also allows you to perform operations on your Koyeb resources directly from your GitHub Actions worflows.

The v2.0 of the Koyeb CLI (Command Line Interface) is now available and ready to let you manage all your Koyeb resources directly from the terminal!

The Koyeb CLI is a critical tool to improve the deployment experience. It provides a fast way to interact with Koyeb when you develop, build and deploy your applications.

You can now easily interact programmatically with the Koyeb platform using the Koyeb Go SDK. The Koyeb Go client allows you to leverage all the capabilities of the Koyeb platfom in your Go applications.

The API Documentation was completely redesigned and updated to include new endpoints. Deprecated endpoints have been removed. We also updated the introduction of the API documentation with clearer instructions on how to get started using the Koyeb API.

Deploying from an Azure Container registry required manually creating a new Secret to store the ACR configuration.

We streamlined the process by adding Azure Container Registry to the list of registry providers on the service creation page, so you can now easily deploy Docker images from Azure Container Registry.

Our edge network was not allowing you to use server-sent events in your application.

Server-sent events are now natively supported on all applications.

Our built-in Service Mesh & Discovery is now fully operational. All the services inside an App can now communicate securely using the Koyeb mesh network.

No configuration is needed to benefit from the Service Mesh & Discovery features. Simply use the automatic private DNS entry of a service to connect and communicate with it.

Previously, the containers hosting your services started even if the network was not ready. This sometimes caused outgoing connections to fail at boot time when the network took more than 100ms to come up.

The issue is now solved and your service code is started after the mesh network is ready.

The Koyeb Edge and Mesh network serve custom 404, 502, and 503 pages to help you identify the origin of errors during deployment or when a runtime failure occurs.

We improved the content of these pages to simplify troubleshooting and improve your deployment experience. Let us know how we can further improve them for you!

Logs of services with horizontal scaling enabled were only displayed for the first node.

The logging page on the Koyeb control panel now lets you track all nodes logs when you scale horizontally.

Previously, deploying from a GitLab private registry was possible but required to manually create a secret with a JSON containing the credentials. This was error-prone and often led to deployment failures.

We simplified this process and added GitLab directly to the list of providers on the service creation page.

To help you get started in minutes and deploy your first application on the Koyeb platform, we now provide a demo application you can deploy in one click.

The application automatically renders a page with your latency to the Koyeb Edge and Core Location where your service is running. It also displays the environment variables available for the service and the HTTP request headers to help you configure your app.

Share your edge latency with us on Twitter and spread the word about Koyeb! 😇