Runex

Feature · Scale & data

Autoscaling that follows your traffic.

Autoscaling adds capacity when traffic rises and removes it when traffic falls. Runex scales your app horizontally with more container replicas, or vertically with more CPU and memory, within the limits you set.

Start deploying →How it works
  • Horizontal replicas
  • Vertical CPU & memory
  • Min and max you control
acme/api · scaling
replicas
2 → 6
trigger
CPU above target
min / max
2 / 8
per replica
CPU & memory limits
status
scaling up
autoscalingrunex

Definition

What is autoscaling?

Autoscaling is the automatic adjustment of an app's compute resources based on demand.

Without it, you size servers for your busiest hour and pay for idle capacity the rest of the day, or size for normal traffic and fall over during a spike.

On Runex, you set a minimum and maximum number of replicas. When load crosses your target, Runex starts more isolated containers and spreads requests across them; when load drops, it scales back down.

Autoscaling works best for stateless apps. Keep sessions and uploads in a database or object store, not in one container's memory or disk.

How it works

How does autoscaling work on Runex?

Set the bounds; Runex moves within them.

  1. You

    Set min and max replicas

    Choose the floor and ceiling for your app.

  2. Runex

    Watch the load

    CPU and memory are tracked per replica.

  3. Runex

    Add replicas

    New containers start and share the traffic.

  4. Runex

    Scale back down

    Extra replicas stop when traffic falls.

Benefits

Why use autoscaling?

Handle the spike, skip the idle capacity.

Survive traffic spikes

Launch days and viral posts don't take you down.

Right-size automatically

No running peak capacity all day.

Hard limits you control

A max replica count caps resources.

More resilience

With several replicas, one crash isn't an outage.

Compare

Horizontal vs vertical scaling: what's the difference?

Horizontal adds more copies; vertical makes one copy bigger. Runex supports both.

Horizontal vs vertical scaling: what's the difference?
Horizontal scalingVertical scaling
HowMore replicas of your containerMore CPU and memory for one container
LimitYour max replica settingThe largest size available
NeedsA stateless appNo code changes
Best forWeb traffic and APIsMemory-heavy, single-process work

Example

Example: a product launch spike

An API goes from a normal day to a launch-day spike without anyone getting paged.

  1. The API runs on 2 replicas on a normal day
  2. The launch post goes live; CPU climbs past target
  3. Runex adds replicas up to the max of 8
  4. Requests spread across every replica
  5. Traffic settles in the evening
  6. Runex scales back down to 2 replicas
autoscaling
morning
2 replicas
launch
8 replicas · max reached
errors
none
evening
2 replicas
status
steady
illustrative namesrunex

FAQ

Questions developers ask

Short answers about autoscaling. More in the documentation.

What is the difference between autoscaling and load balancing?

Load balancing spreads requests across the replicas you have. Autoscaling changes how many replicas you have. Runex does both together.

Does autoscaling scale my database too?

No, autoscaling adds app replicas, not database capacity. Size your database for peak connections and use connection pooling when many replicas connect at once.

What is a replica?

An identical running copy of your app. On Runex, each replica runs in its own isolated container with its own CPU and memory limits.

Why isn't my app a good fit for horizontal scaling?

If it stores sessions or uploads on local disk or in memory, replicas won't share that state. Move it to a database, cache, or object storage first.

Next step

Ship with autoscaling today.

Sign up, install the GitHub App, and deploy your first app. It's free to start.

Deploy with Runex →Getting started guide