Skip to content

Create a Migration Plan

The plan creation wizard guides you through a series of steps that capture everything the suite needs to generate networking guidance, tool recommendations, cutover strategies, and downloadable assets. Each step builds on the previous one, so the suite can tailor its output to your specific environment.

  1. Sign in to the Kafka Mobility Intelligence suite.

  2. Navigate to the Dashboard.

  3. Select the Migration Plans tab.

  4. Click New Plan.

The wizard opens with the first step. Complete each section below in order.


Specify the source and destination environments for the migration.

  • Source Selection: Choose the source cluster type and authentication mechanism.
  • Destination Selection: Choose the destination cluster type.

Supported source cluster types include:

Source TypeNotes
Confluent PlatformOn-premises Confluent deployment
Confluent CloudConfluent-managed cloud
Amazon MSK (SCRAM or IAM)AWS managed Kafka
Apache Kafka (open-source)Self-managed Kafka cluster

Supported destination cluster types include Confluent Cloud, Confluent Platform, and WarpStream.

Cloudera is not a separate source cluster type. To migrate from a Cloudera environment, select the matching Kafka source type and register a Cloudera Schema Registry in Pre-Migration Setup; the suite then uses the Tern engine for schema migration.

The suite uses these selections to architect the migration bridge, filter compatible connectivity methods, and determine which security protocols apply.

Cluster type selection in the plan wizard


When the suite detects an AWS-based source cluster, it offers automated provisioning through KCP (Kafka Cloud Provisioner).

KCP provides two key capabilities:

  • Automated Discovery: Scans your AWS environment for MSK configurations, including topics, schemas, and networking details.
  • Target Infrastructure Provisioning: Auto-provisions Confluent Cloud infrastructure to match the source configuration.

KCP automated provisioning options


Define what assets to include in the migration.

Asset types:

  • Topics
  • Schemas
  • Consumer Groups
  • ACLs

Environment category:

CategoryCharacteristics
ProductionStrict SLAs, zero-downtime requirements, comprehensive monitoring
Non-ProductionRelaxed constraints, faster testing cycles, cost-efficient resource allocation

Migration scope selection


This step appears when the migration involves AWS MSK discovery or KCP provisioning.

Credential Type: Choose from Access Key, Assume Role, or Environment credentials.

Provide the following details as applicable:

  • AWS Account ID
  • Secret Access Key
  • AWS Regions where source MSK clusters are located

AWS credentials configuration


Define the networking profile for the source cluster.

  • Public Accessibility: Toggle whether the source cluster is publicly accessible.
  • Source Network Type: Select from Public Internet, VPC Peering, PrivateLink, Transit Gateway, Private Network Interface, Private Service Connect, VPN/Direct Connect, or Direct Connect.
  • Network Restrictions: Indicate any firewalls, IP whitelisting rules, NAT gateways, port restrictions, or proxies that affect connectivity.

Source cluster networking configuration


Configure the destination cluster networking. This section mirrors the source cluster configuration.

  • Choose an existing cluster, or select Destination cluster is not yet set up to use KCP provisioning.
  • Specify the Cloud Provider and Region.
  • Configure networking: public access toggle, Network Type (Public Internet, VPC Peering, PrivateLink, Transit Gateway, Private Network Interface, Private Service Connect, VPN/Direct Connect, Direct Connect), and any restrictions.

Target cluster networking configuration


The suite evaluates the network profiles captured in the source and target steps and presents viable connection methods.

  • Connection methods: Public Internet with TLS, VPC Peering, PrivateLink, Transit Gateway, Cross-Region Peering, VPN Overlay, Dedicated Interconnect, or VPN/Direct Connect.
  • Automated warnings: The suite flags cross-cloud migration concerns such as egress costs and latency implications.

Inter-cluster connectivity evaluation


Choose a migration strategy that determines how applications are redirected during cutover.

Manual cutover approach. Applications must be individually reconfigured to point to the new cluster after migration completes.

Gateway strategy selection


When Gateway-Assisted Migration is selected, the next step asks you to Select a Gateway Resource: pick a saved gateway to use for this migration, or create one from here. Once you pick one, the step shows a summary card for that gateway with a View in builder link that opens it in a new tab, so review it there before you finish the plan.

The plan does not ask you to configure the gateway again. It reads the gateway you select and derives the rest from its Init state, including the authentication mode, the broker identification strategy, and the custom domain. Those three shape the checklist on the plan’s Gateway tab: a swapping gateway gets an extra JAAS preparation item, and a host/SNI gateway gets a DNS wildcard item plus its own domain in the bootstrap endpoint.


If schema migration is included in the migration scope, this step captures planning details that inform the suite’s recommendations for schema handling.

Source Message Serialization: Select how messages on the source cluster are serialized.

OptionDescription
Confluent-compatibleStandard Confluent wire format (magic byte 0x00). No transform needed on the destination.
Cloudera-nativeCloudera’s proprietary serde format. The wizard additionally asks for the specific serde protocol (0, 1, 2, or 3).
Raw / NoneMessages are not schema-serialized (plain bytes or JSON without a schema registry). No transform needed.

When Cloudera-native is selected, the wizard displays a follow-up prompt for the Cloudera serde protocol:

ProtocolWire format
0Confluent-compatible; no re-framing needed on the destination
1Requires metadata ID and version ID lookup during re-framing
2Pure byte re-framing with no registry lookup
3Pure byte re-framing with no registry lookup

When the source is Cloudera-native on protocol 1, 2, or 3 and the recommended tool is MirrorMaker 2 or Replicator, the plan generates a recommended re-frame SMT (Single Message Transform) template and a downloadable sample SMT Java source. See the Assets tab in View a Migration Plan.

Schema registry planning details


Based on your answers throughout the wizard, the suite recommends a migration tool.

The recommendation includes:

  • Recommended tool: The best-fit migration tool for your environment.
  • Rationale: Why this tool is recommended given your source, target, and scope.
  • Alternative tools: Other viable options.
  • Trade-offs: Key differences between the recommended and alternative tools, including rollback characteristics.

Rollback / Failback Requirement: This step also asks whether the migration must support rollback or failback after cutover.

AnswerEffect on recommendation
No (default)Tool recommendation is unchanged from today’s behavior; source type and Confluent Cloud tier drive the choice.
YesCluster Link is excluded from the recommendation and from the alternatives list when the source is non-Confluent (open-source Apache Kafka, Amazon MSK). A cluster link destination must always be a Confluent cluster; a non-Confluent source cannot be a rollback target through Cluster Link. MirrorMaker 2 or Replicator is recommended instead, as both support reverse replication to any Kafka cluster.

When rollback is required and the source is Confluent (Confluent Platform or Confluent Cloud), Cluster Link remains eligible because the reverse link destination is still Confluent.

WarpStream Orbit: The rollback question does not change the recommendation for WarpStream targets, where Orbit is always recommended, but the plan surfaces that Orbit promotion is irreversible with no rollback path.

Tool preference recommendation


Define how the cutover from source to destination will be executed.

Strategy types:

StrategyDescription
Big BangSimultaneous switch of all workloads at a single point in time
RollingIncremental migration in waves, moving workloads gradually
Gateway AssistedNetworking-layer shift managed by the gateway, transparent to applications

Delivery Guarantee:

GuaranteeBehavior
At-Least-OnceDuplicates are acceptable; simplest to implement
Exactly-OnceClean cutover with coordinated downtime; no duplicates or loss
At-Most-OnceSome data loss is acceptable; fastest cutover
Zero-Producer-DowntimeProducers experience no interruption during cutover; mirroring continues until after producer switch-over

Cutover planning configuration


Provide an overview of the applications that interact with Kafka topics on the source cluster.

  • Consumer Groups: Applications consuming from source topics.
  • Producers: Applications producing to source topics.

Application dependencies overview


Identify any Kafka Connect usage that is relevant to the migration plan.

Connector identification


The final step presents a summary of all planning inputs collected throughout the wizard. Review each section to confirm accuracy.

Click Create to finalize the plan. The completed plan becomes available from the Migration Plans tab on the Dashboard, where you can view its details, download assets, or create a migration job.

Plan review and creation summary