How dockxo is built, and how to use it well.
Design notes on the routing layer, the database cluster and the backup format, plus practical guides. New posts land with every release.
dockxo vs Virtuozzo Application Platform (Jelastic): a PaaS on your own VPSes, without the licence
Virtuozzo Application Platform, formerly Jelastic, is licensed per cluster from $192 a month and needs a rack of infrastructure nodes before the first app runs. dockxo does the everyday PaaS jobs on the servers you already rent, free during the beta. A fair, numbers-first comparison.
Introducing dockxo: deploy to your own servers, free and unlimited during the beta
dockxo is a deployment panel for people who own their servers. Today it opens as a public beta: every feature, every limit off, no card.
How dockxo routes traffic: one Traefik per server, generated from what is running
A tour of the routing layer: why every server runs its own Traefik with a file provider, how routes are regenerated from Docker events, and how load-balancer members hide behind :8099.
Zero-downtime deploys with rolling swaps
What happens between a git push and the moment the new container takes over, and why the old one stays until the new one answers on its port.
Managed MariaDB: primaries, replicas on other servers, and a router in front
How the managed database kind grew from a single container into a cluster with GTID replicas, MaxScale read/write split and automatic failover, without leaving the panel.
Backups that survive the panel: the sealed project file
Why every backup run starts by writing a sealed copy of the project's configuration to your bucket, and how that makes a bucket enough to rebuild everything on a new panel.
One domain, many servers: load balancing across cheap VPSes
How a dockxo load balancer spreads one domain over the same service on several servers, why members hide behind :8099, and how sticky sessions reach the exact container.
Uptime monitors with Telegram and email alerts, built into the panel
The Monitors page checks any site every 20 seconds to an hour, explains why it is down, and tells your Telegram chat or your inbox. Plus service notifications for failed deploys and crashed containers.