Service Instance Scaling on Linux

This guide explains how to run extra instances of MetaDefender Storage Security (MDSS) services so your deployment can handle more work in parallel. It's an optional setting for larger or busier environments.

If you've read the Scan Priorities & Throughput Tuning guide, here's the key difference:

• That guide tunes how hard a single service container works (more workers, bigger batches).
• This guide adds more containers of the same service, so the work is shared across them.

How it works

By default, each scalable service runs as one container. You can add up to two extra containers per service using the *_ADDITIONAL_INSTANCES settings.

The value is simply how many extra containers to add on top of the default one:

The maximum is 2 additional (3 total). Any higher value is automatically reduced to 2.

When MDSS starts, it reads these values and launches the extra containers automatically. They join the same workload and start sharing it immediately — no other configuration is needed.

What you can scale

There are two groups of services you can scale.

1. Central services

These are the core engines of the scanning pipeline:

2. Storage connectors (per integration)

Each storage type you connect to has its own connector container. You can scale each one independently, so you can add capacity only where you need it:

Examples

Example 1 — Default (no scaling)

xxxxxxxxxx

SCANNING_SERVICE_ADDITIONAL_INSTANCES=0

One scanning container runs. This is the default and is right for most deployments.

Example 2 — Add one extra scanning container

xxxxxxxxxx

SCANNING_SERVICE_ADDITIONAL_INSTANCES=1

Two scanning containers run and share the scanning workload — roughly doubling scanning capacity.

Example 3 — Scale the whole central pipeline

xxxxxxxxxx

DISCOVERY_SERVICE_ADDITIONAL_INSTANCES=1

SCANNING_SERVICE_ADDITIONAL_INSTANCES=2

REMEDIATIONS_SERVICE_ADDITIONAL_INSTANCES=1

Discovery runs 2 containers, Scanning runs 3, and Remediation runs 2 — useful for a large, busy environment where every stage needs more throughput.

When should you scale instances?

Add instances when a single container is already working at full capacity and your server still has spare CPU and memory to run more containers.

A good order to follow:

1. First, tune the single container using the prefetch and consumer-multiplier settings (see Scan Priorities & Throughput Tuning). This is the simplest change and often enough.
2. Then, add instances if one container — even fully tuned — still can't keep up, and you have the hardware to run more.

If your server's CPU is already near 100%, adding more containers will not help — they'll compete for the same resources. In that case, the answer is more hardware (a bigger server, or additional servers), not more instances.

How to apply changes

1. Set the desired ADDITIONAL _INSTANCES values in your environment configuration (_customer.env*]
2. Restart MDSS so the extra containers are created

Things to keep in mind

• The default (0) is right for most deployments. Only scale when you have a clear throughput need and spare hardware.
• The maximum is 2 additional instances (3 total) per service. Higher values are automatically capped.
• Only the services listed in this guide can be scaled. Other MDSS services are designed to run as a single instance and should not be duplicated.
• More instances need more resources. Each extra container uses additional CPU and memory. Make sure your server has the capacity before scaling.
• Instance scaling and worker tuning work together. You can both add instances and tune each container's workers/prefetch for maximum throughput. </content>