Redis

AddisDB gives Redis a proper editor instead of a bare terminal. You type real Redis commands, and replies are rendered as a grid: an array becomes one row per element, a hash becomes a single keyed row, and a scalar becomes one cell.

Who it is for

Redis keeps everything in memory, which makes it the default answer for anything that must be fast and can be rebuilt: caches, sessions, rate limiters, job queues, leaderboards, ephemeral counters. Its data structures — lists, sets, sorted sets, streams — do real work that would otherwise be application code.

It is also single-threaded for command execution, which is a feature and a trap at once. Every command is atomic without you asking, and one expensive command blocks every other client until it finishes. That is why KEYS is a production incident and SCAN is not.

Use it alongside your primary database, not instead of one. It is at its best holding data you would be comfortable losing.

Set up the server

1. Managed: Redis Cloud, Amazon ElastiCache, Azure Cache for Redis or Upstash. Copy the endpoint host, port and password from the console.
2. ElastiCache lives inside a VPC and is not reachable from the internet — you will need a bastion host (see the tunnel step below).
3. Local: docker run -p 6379:6379 -d redis, or brew install redis.
4. Redis 6+ supports usernames via ACLs; older servers use just a password. The default username is default.
5. For read-only access, create an ACL user restricted to read commands rather than handing out the default password.

# A read-only ACL user, all keys, read commands only
ACL SETUSER addisdb on >a-strong-password ~* +@read +info +client|list

Connect from AddisDB

1. New Connection → Redis under Key–Value. Port prefills to 6379 and username to default.
2. Enter the host and password. Set Database index to the numeric database you want — 0 unless you know otherwise.
3. For ElastiCache or any VPC-only server, open Tunnel (optional) and add your bastion host. AddisDB opens the SSH tunnel on connect and closes it on disconnect.
4. Test, then Save.

Redis Cluster splits the keyspace across nodes and answers a request for the wrong shard with a MOVED redirection. A single-node or proxied endpoint — which is what most managed offerings hand you — avoids the question entirely.

How to query it

GET session:8f21
HGETALL user:42
LRANGE jobs:pending 0 -1
ZREVRANGE leaderboard 0 9 WITHSCORES
SCAN 0 MATCH cache:* COUNT 100

SCAN is a cursor, not a listing: it returns a new cursor and a batch of keys, and you feed the cursor back in until it comes home as 0. It is the safe way to walk a keyspace, because it never blocks the server the way KEYS * does.

When you do not know what a key holds, TYPE tells you, and OBJECT ENCODING tells you how Redis is storing it — which is often the explanation for surprising memory use. MEMORY USAGE on a suspect key answers the rest.

TYPE user:42
OBJECT ENCODING user:42
MEMORY USAGE user:42
TTL session:8f21

What AddisDB gives you

• Replies rendered as a grid rather than raw protocol output, with the JSON tab for the exact structure.
• Read-only enforcement written for Redis — mutating verbs like SET, DEL and HSET are blocked on a read-only connection.
• Query tabs and history, so the command you worked out last week is still there.
• SSH and AWS SSM tunnels for reaching a server inside a private network.
• The Live Monitor over CLIENT LIST and INFO, including blocked clients and per-database key counts.