Find records in a topic

The Data tab of a topic shows its records. You need the records privilege on the cluster to see it.

Pick the records to showCopy Anchor Link

1. Open a topic from Topics. The Data tab shows the newest records first.
2. To read from the start of each partition, select Oldest in the mode switch. To return to the end, select Newest.
3. To search, type in Search key or value…. klens keeps the records whose key or value holds the text, ignoring case.
4. To read some partitions only, select Filter, then Partition, and check the partitions. Each partition shows how many records it holds.
5. To read a time range, select Filter, then Timestamp, and pick a range such as Last hour. For other times, select Custom range… and set From and To.

klens searches on the server and reads more records as you scroll. A search on a large topic can stop at the read timeout before it reads every offset. The page then shows Partial scan. Scroll down to keep reading.

The search, the partitions, and the schema stay in place when you change the mode. The time range resets when you switch to Live and back.

Decode values without a schema IDCopy Anchor Link

klens decodes Avro, JSON Schema, and Protobuf values that carry a Schema Registry ID. A value written without that ID shows as raw text. To decode such values, open Decode with schema… and select a subject. The Suggested group lists <topic>-value. To show raw values again, select Raw.

The picker appears only when the cluster has a schema_registry and a loaded value has no schema ID.

Open one recordCopy Anchor Link

• To open a record from the list, select its row. A side sheet shows the partition, offset, size, timestamp, schema, key, value, and headers.
• To open a record by its address, select Go to offset, enter the Partition and the Offset, and select Open record. klens reads the record from Kafka, even when it is not in the loaded list.

In the sheet, Pretty formats a JSON value, and Raw shows it as stored. Download value saves the value as <topic>-<partition>-<offset>.json.

When the partition holds no record at the offset, the sheet shows Record not found. Retention or compaction may have removed the record, or the offset may hold a transaction marker.

Share a recordCopy Anchor Link

The address of an open record is /cluster/<cluster>/topics/<topic>?partition=<p>&offset=<o>. To copy it, select Copy link to this record in the sheet. The link opens the Data tab with the record open.

Download the records that matchCopy Anchor Link

To save every record that matches the current view, select Download NDJSON. klens writes one JSON record a line to <topic>.ndjson. The file holds every match in the partitions, the time range, and the search you picked, in the current order. Records that arrive after the download starts are not in the file.

The download is not available in Live mode, or when the partition filter excludes every partition.

Read obfuscated topicsCopy Anchor Link

An Obfuscated badge means an obfuscation rule covers the topic. Hidden fields show as *** or as a kx: token, and equal values get equal tokens. The search matches the hidden text, not the original value.