Scale with your traffic automatically
Your APIs, websites and server fleets grow and shrink with demand. Traffic is shared across healthy instances, and new versions go live without dropping a request.
Capabilities
What you get
Autoscaling on CPU and memory
Set a minimum and maximum instance count per region and the CPU and memory thresholds that should trigger scaling. Available for API, Website and master-worker Server deployments.
Round-robin load balancing
The platform’s managed reverse proxy sends each request to the next healthy instance. Included at no extra cost for APIs, websites and secure WebSocket servers.
Health-aware routing
Instances that fail their health check are taken out of rotation automatically and put back when they recover.
Zero-downtime deploys
New versions start alongside the current ones and receive traffic only once healthy. The previous version is kept for rollback.
Graceful server rollouts
For game and real-time servers, old servers stop taking new sessions and drain for a grace period so connected users are not cut off.
Plan-based limits
Each plan sets how many instances you can run per region. Larger plans allow larger fleets.
How autoscaling decides
Measure
The platform samples CPU and memory for every instance continuously.
Scale up
When average usage stays above your scale-up threshold for a few minutes, an instance is added (up to your maximum).
Scale down
When usage stays below your scale-down threshold for a longer period, an instance is removed (never below your minimum).
Stay stable
Scale-down deliberately waits longer than scale-up, so short dips do not cause instances to flap on and off.
NOTE: Every running instance is charged by the minute. Your maximum instance count is also your cost ceiling set it with your budget in mind.
How load balancing work
01
Request arrives
The managed reverse proxy receives the HTTPS request for your API or website.
02
Next healthy instance
The proxy picks the next healthy instance in rotation — A, then B, then C, then A again.
03
Response returned
The instance handles the request and the response goes back to the caller.
04
Failures routed around
If instance B becomes unhealthy, requests flow only to A and C until B recovers.
Frequently Asked Questions
Everything you need to know about MultiConnect Cloud.
It depends on your plan. The limit is shown on every plan card on the Pricing page.
No. It is included in every API, website and server plan. You pay only for the instances you run.
No. Databases have a fixed size set by their plan, plus a standby or replica for failover. Change plan from the dashboard when you need more capacity.
No. Combined mode runs one server per region. Use master-worker mode if you need an autoscaling fleet.