Streams live viewer

The Streams panel tails a table’s DynamoDB Stream and renders each change event live as it happens.

Enabling a stream first

You can only tail a stream that exists. Open table settings, enable DynamoDB Streams, and choose a view type:

View type	What each record carries
NEW_AND_OLD_IMAGES	both the before and after images
NEW_IMAGE	the item after the change
OLD_IMAGE	the item before the change
KEYS_ONLY	only the key attributes

For the richest live view (seeing both old and new images), pick NEW_AND_OLD_IMAGES. Apply, then open the Streams panel.

Opening the viewer

Right-click the table and choose Streams. Tablyne resolves the table’s LatestStreamArn via DescribeTable. If the table has no stream enabled, the panel says so — go back to table settings and enable one.

Once resolved, the viewer starts polling automatically. New events appear at the top of the list.

Reading the feed

Each record is rendered with its event type, the item keys, and the relevant image:

• INSERT — green; shows the new image.
• MODIFY — cyan; shows the new image.
• REMOVE — magenta; shows the old image with a strike-through.

INSERT   {"pk":"USER#42","sk":"PROFILE"}
{
  "pk": "USER#42",
  "sk": "PROFILE",
  "name": "Ada"
}

The color coding lets you scan a busy feed at a glance. Images are pretty-printed JSON (converted from DynamoDB’s typed attribute format).

Controls

• Pause / Resume — stops and restarts polling. The live dot turns grey when paused.
• Clear — empties the on-screen list without affecting the stream.
• Minimize — collapses the panel into a floating chip (see below).

A live indicator pulses while polling. Tablyne polls every few seconds and keeps the most recent 200 records in view; older ones scroll off.

Background multi-tail

Minimizing a stream doesn’t stop it. The panel stays mounted as a small floating chip in the corner — showing the table name, a live/paused dot, and a running event count. Polling continues in the background while you work elsewhere in the workspace.

You can minimize several streams at once; chips stack so you can tail multiple tables simultaneously. Click a chip to restore its full panel, or the ✕ on the chip to close that tail.

Limitations (be honest)

The live viewer is intentionally a monitoring tool, not a stream-processing engine:

• Single shard. Tablyne reads the first shard of the stream. High-throughput tables that fan out across many shards will only show events from that one shard.
• LATEST only — no backfill. The iterator starts at LATEST, so you see events that occur after you open the panel. There is no history replay; events written before you started tailing are not shown.
• No filtering or persistence. The feed is in-memory and capped at 200 records; closing the panel discards it.

These trade-offs keep the viewer fast and simple. For full, durable stream processing, wire the stream to a Lambda or Kinesis consumer.

Cloud entitlement

Tailing a stream on AWS cloud requires an active license or trial — see licensing. Against DynamoDB Local it works without a license.

Related

• Table settings — enable streams and choose the view type.
• Backups & PITR — snapshot data instead of watching it change.
• Data grid — browse the current table contents.
• Concepts and patterns: the streams learn guide.