Database Bubbles

Cockpit's Console panel auto-detects database connection strings and opens a database bubble — a built-in client with a schema browser, a data browser, and a query window, scoped to that one database. No more flipping out to psql / mysql / DataGrip / RedisInsight just to peek at a row.

Cockpit recognises four database flavours today:

The Console input bar parses the URI on paste and opens the right bubble automatically — see Command Input for the full set of triggers.

PostgreSQL

A PostgreSQL bubble gives you a connection, a schema browser, and a query window — no need to leave Cockpit for quick lookups.

Open one by pasting a connection string into the Console input bar:

postgresql://user:password@localhost:5432/mydb

postgres://... also works.

Layout

The bubble has a schema browser on the left and three tabs on the right:

• Left sidebar: schema tree — schemas → tables and views → columns. Filter box at the top; click a table to select it.
• Right "Structure" tab — the structure of the selected table: columns (type, nullability, default), primary key (🔑 marker), foreign keys (target table / column), indexes (type, fields). This is where PK/FK/indexes live, not in the schema tree itself.
• Right "Data" tab — the table data, paginated (50 rows per page by default).
• Right "SQL" tab — query window. Type SQL, press Cmd/Ctrl+Enter or click Run, results render as a table below.

What the Data tab can do

• Click a column header to sort ascending / descending.
• Filter per column with familiar operators: =, !=, >, <, >=, <=, LIKE, NOT LIKE, IN, IS NULL, IS NOT NULL.
• Hover a cell to see the full content (useful when a long value is truncated; 350ms delay before the tooltip appears).
• Inline edit — double-click a cell to change a value, save or cancel.
• Add new row — form to insert.
• Multi-select + delete — tick rows and confirm.
• Export — copy or download selected rows or the whole table as CSV / JSON.
• Query timing — shown next to the result count as xx ms.

Common issues

• Connection refused — the database isn't reachable from your machine. Check the host / port, your VPN if you need one, and pg_hba.conf if it's your own server.
• Password authentication failed — wrong credentials in the connection string, or the user doesn't have permission to connect from your IP. URL-encode special characters in the password (@ → %40, etc.).
• Long queries hang — Cockpit doesn't auto-cancel long queries. Close the bubble to kill the connection, or run SELECT pg_cancel_backend(pid) from another bubble.

MySQL

The MySQL bubble is structurally near-identical to the PostgreSQL bubble: same left-hand schema tree, same Structure / Data / SQL three tabs, same pagination (50 rows/page), same sorting and filtering, same inline editing and export.

Open one by pasting a connection string:

mysql://user:password@localhost:3306/mydb

Implementation differences from PostgreSQL

Behaviour is the same on the surface; SQL dialect details differ under the hood (mostly invisible to you):

• Identifiers quoted with backticks ` instead of ".
• Row counting uses COUNT(*) AS cnt, not PostgreSQL's ::int cast.
• Deletion uses DELETE … LIMIT 1, not PostgreSQL's ctid subquery.

Common issues

• Access denied — wrong credentials, or the user can't connect from your IP. Check mysql.user host grants on the server. URL-encode special characters in the password.
• Unknown database — the database name in the connection string doesn't exist on the server.
• TLS handshake failures — follows standard URI parameters for TLS configuration; if you're connecting to a managed service that requires TLS, check the provider's docs for the right query string.

Redis

A Redis bubble gives you a key browser, an INFO viewer, and a CLI tab for raw commands. Open one by pasting a connection string:

redis://localhost:6379
redis://:password@host:6379/2
rediss://host:6380       # TLS

The URI follows the standard redis://[:password@]host:port[/db] form. Use rediss:// (extra s) for TLS. The bubble's title bar shows Redis version, key count, and memory usage.

The bubble has three tabs across the top:

1. Data tab (key browser)

The default view. You get a list of keys with a filter at the top that takes a Redis pattern (*, ?, [abc]). When the database has many keys, a Load more button at the bottom pages through them. Click a key to see its value.

Each key shows:

• Type badge — STRING, HASH, LIST, SET, ZSET, STREAM — coloured so you can scan a list quickly.
• Value preview — the value for STRING, the field/value pairs for HASH, the items for LIST / SET / ZSET, etc.
• TTL — formatted as seconds, minutes, or hours (60s, 2m30s, 1h5m).

Long values are truncated in the preview; hover or click in for the full content.

2. Info tab

Shows the output of the INFO command, parsed into sections — replication, memory, connected clients, replication state, and so on. Handy for quick triage without switching to the CLI.

3. CLI tab

A raw Redis command line. Supports space-separated arguments and quoted strings. Type any command and hit Enter:

INFO replication
DBSIZE
SCAN 0 MATCH user:* COUNT 100

Results are formatted nicely — arrays come back as numbered lists, integers as integers, (nil) as (nil).

Common issues

• WRONGPASS — password missing or wrong. Include it in the URI: redis://:yourpassword@host:port.
• Connection timeout — Redis isn't reachable from your machine. Check the host / port; if it's a managed service, verify it allows connections from your IP.
• ERR DB index is out of range — the /db portion of the URI references a database number that doesn't exist (Redis defaults to 16 databases, numbered 0–15).

Neo4j

A Neo4j bubble lets you explore a graph database — browse the schema, run Cypher queries, and see results as an actual interactive graph, not just rows of records.

Open one by pasting a connection string. Neo4j supports four URI schemes; pick whichever your server documentation tells you to use:

neo4j://localhost:7687
neo4j+s://yourdb.databases.neo4j.io       # TLS
bolt://localhost:7687
bolt+s://yourdb.databases.neo4j.io        # TLS

The +s variants enable TLS. neo4j:// is the modern scheme (routing-aware); bolt:// is the older direct-connection scheme.

The bubble has three tabs.

1. Schema

A read-only overview of the database structure (counts refresh after queries run):

• Labels — every node label and how many nodes have it. Click a label to fire a query that pulls nodes of that type.
• Relationship types — every relationship type and its count. Click a type to pull relationships of that kind.
• Property keys — all property names used anywhere (listed, not clickable).
• Indexes — name, type, fields, state.
• Constraints — uniqueness and other constraints.

Use this when you sit down to a database for the first time and need to know what's in it.

2. Cypher

A query window. Press Cmd/Ctrl+Enter to run:

MATCH (p:Person)-[:KNOWS]->(friend)
WHERE p.name = 'Alice'
RETURN p, friend

Results render as a table at the bottom — one row per record, with columns for each returned variable.

3. Graph

When your Cypher returns nodes and relationships, switch to this tab to see them as an interactive force-directed graph:

• Nodes are circles, coloured by label.
• Relationships are arrows, labelled with their type (e.g. KNOWS).
• Drag nodes around; the layout settles itself. Node positions scale proportionally when the bubble is resized.
• Click anything to see its properties.

If your query only returns scalar values (e.g. RETURN count(*)), the Graph tab is empty — use the Cypher tab's table view instead.

Common issues

• Unauthorized — wrong credentials in the URI. Format: neo4j://user:password@host:7687.
• Connection refused — Neo4j isn't running or isn't reachable. Default Bolt port is 7687 (separate from the HTTP 7474 you might see in browser docs).
• Slow queries on big graphs — add LIMIT while exploring. The Graph tab tries to render everything you return; a 100k-node result will lock up the layout.