Improve Your Producers Efficiency With This Powerful Calculator
Apache Kafka
5
min read

Cut AWS MSK Costs and Boost Performance: Migrate to Graviton

From Intel to Graviton: How Superstream Enables Safe, Seamless AWS MSK Migrations
Idan Asulin

Migrating from Intel-based instances to AWS Graviton-based instances is one of the most effective ways to cut infrastructure costs and improve performance. But for teams running Amazon MSK (Managed Streaming for Apache Kafka), the migration path is anything but straightforward. Concerns around efforts, compatibility, stability, and downtime often stall these initiatives.

Superstream was built to make complex Kafka infrastructure optimization as simple as possible—and this includes helping customers migrate their MSK workloads to Graviton with zero risk and minimal effort.

Why Graviton Matters for MSK Customers

AWS Graviton processors, built on ARM architecture, offer significant advantages over x86-based Intel instances:

  • Up to 24% price/ performance improvement
  • Up to 29% higher write and read throughput
  • Up to 27% reduction in CPU usage over comparable MSK clusters running on M5 instances.

But these benefits are often out of reach for MSK users due to time and efforts.

Enter Superstream: A Safer, Easier, And Automated Path to Graviton

Superstream is a drop-in optimization layer for Kafka workloads that enables frictionless Kafka configuration, infrastructure, and cost-efficiency improvement.

Here's how Superstream makes the Graviton migration easy:

  • Add an IAM user or role to grant Superstream permission to automatically discover existing AWS MSK clusters.
  • Superstream analyzes the clusters and identifies opportunities for Kafka-level optimizations.
  • Optimization fixes are applied automatically.
  • Configure the Superstream autoscaler to enable seamless migration from Intel-based instances to Graviton.
  • When infrastructure requirements are met, Superstream automatically performs the migration.
  • Step 1: Superstream to auto-discover your AWS MSK clusters

    Step 2-3: Analysis and Optimization identification

    Step 4: Configure Superstream Autoscaler for AWS MSK

    Behind-the-scenes: Rolling Replacement of Brokers

    MSK performs a rolling update, replacing one broker at a time:

    • Step 1: A broker is taken out of service.
    • Step 2: A new broker with the updated instance type is launched.
    • Step 3: The new broker joins the cluster.
    • Step 4: Partition replicas and leadership roles are reassigned as needed.
    • Step 5: The old broker is terminated.

    This process is repeated for each broker in the cluster.

    Availability and Safety

    • Zookeeper remains operational during the change.
    • Write/Read continues—data is not lost.
    • MSK monitors broker health and will retry steps if something fails.
    • The change does not require client-side reconfiguration (brokers retain their DNS endpoints).

    You're just a few clicks away from unlocking significantly more value for every dollar you're spending—it's absolutely worth it.

    The Fastest Way To Optimize
    Your Apache Kafka Costs
    Recover at least 43% of your Kafka costs
    and save over 20 hours of Ops time weekly.
    Start free now
    Free
    $0
    / Savings report
    Growth (Most Popular)
    $10
    / TB in / Optimized Cluster / Month
    • Starts at 50TB
    • Annual plans only
    Enterprise
    Custom
    • Private link
    • SSO
    • Annual plans only
    * Superstream does not ingest your data — we just use this metric to size your cluster and our value to it.
    * Do you purchase through AWS Marketplace? You might be eligible for a discount when subscribing via AWS MP.
    Your Very Own AI Team

    Save Time, Save Money, and Start Today

    Start free
    Built and Trusted by Data Engineers worldwide