Explore topics and records · Application Manual

First-party product documentation. Instructions are verified against the dFlow~IQ desktop implementation. Labels appear exactly as they do in the application where practical.

03.1

Browse topics

Open Workspaces, select a workspace, then choose Kafka → Topics.
Use Search topics to filter the current fetched page.
Select a topic to inspect Partition Count, Replication Factor, Is Internal and Configuration.
Use Browse records for ad hoc inspection or Ask AI after reviewing shared context.

Page sizes are 10, 25, 50 or 100. Refresh after broker-side changes.

03.2

Create a topic

Topic Name

Required broker topic name.

Partitions

Positive integer; default 1.

Replication Factor

Positive integer not exceeding available brokers; default 1.

Retention (ms)

Default 604800000, equivalent to seven days.

Cleanup Policy

delete or compact.

Min In-Sync Replicas

Positive integer; default 1.

Max Message Bytes

Positive integer; default 1048576.

Topic creation and deletion change the cluster. Use the technology guide to understand the operational impact before applying production values.

03.3

Choose where browsing begins

Latest 1,000 per partition

Loads a bounded tail from every partition.

Earliest

Starts at the earliest records still retained.

Offset

Starts from one validated partition and offset.

Timestamp

Starts each partition at its first record at or after the selected local date and time.

Offsets must be inside the broker's retained range. A timestamp with no later record starts at that partition's end.

03.4

Create a saved consumer

Choose Kafka → Consumers → Create consumer.
Enter a unique Consumer Name and choose a Topic.
Select key and value deserializers independently.
For SchemaRegistry, select the required subject.
Choose Latest or Earliest and save.
Select Consume on the saved definition.

Supported deserializers include String, Int32, Int64, SchemaRegistry, Avro(Embedded), Base64, Hex and UUIDBinary.

03.5

Filter the local record index

Key regex or exact key.
A bounded retained-topic scan for one exact serialized key using Search topic.
Header name and optional header value.
Object-field JSON paths such as $.customer.id.
From and To timestamps in the desktop's local timezone.

Each browser run uses a temporary SQLite index bounded to 250,000 records and 512 MiB. It is removed when the browser closes.

03.6

Inspect exact record content

Select a metadata row showing timestamp, key, partition, offset and size.
dFlow~IQ fetches that exact record from Kafka on demand.
Use Fold all or Unfold all for structured content.
Use Copy record content or Download record content for local analysis.

If retention removed the broker record after indexing, the detail pane reports Record is no longer available in Kafka.