Browse Redis keys without KEYS * and without write commands

Browse Redis keys with a SCAN cursor and MATCH filter, preview six value types with TTL, and run an allowlist of read-only commands, directly or over SSH.

Transport: ego-component/eredis; upstream go-redis for TLS or named ACL users (port 6379 by default)

What you can do today

  • Cursor-based key browsing with SCAN, MATCH patterns and Load more keys
  • Key type and TTL for every key you open
  • Value previews for strings, hashes, lists, sets, sorted sets and streams
  • Read-only commands such as GET, MGET, HGET, LRANGE, XRANGE, TTL, SCAN, DBSIZE and INFO
  • Database index selection, named ACL users and TLS with certificate verification
  • SSH tunnels with password, private key or agent authentication

Not supported yet

  • Redis Cluster and Sentinel profiles
  • Editing keys, values or TTLs
  • Full pagination of large collection values inside the key preview

Look inside Redis without risking it

Redis is easy to damage from a console: KEYS * on a large instance blocks every client, and a mistyped DEL or FLUSHDB cannot be undone. DevOpsPad’s Redis connector is built so neither can happen from the app.

The Explorer walks the keyspace with SCAN, never KEYS. Type a MATCH pattern such as session:* to narrow it, and use Load more keys to continue from the current cursor. Open a key to see its type, its TTL and a bounded preview of its value: up to 100 collection entries, or 64 KiB of string content.

Read-only commands in the editor

The query editor accepts plain commands or a JSON array when a key contains spaces:

SCAN 0 MATCH session:* COUNT 100
["GET", "a key with spaces"]
HGET user:1042 email

Commands are checked against a read-only allowlist before they reach the server. Responses over 8 MiB are rejected with a hint to use a bounded range or a smaller SCAN batch, so a huge list does not freeze the window.

ACL users, TLS and private networks

Set a username for Redis 6+ ACL users, or leave the default user. Enable Use TLS with certificate verification for managed Redis endpoints. For those two modes DevOpsPad uses the upstream go-redis client directly so named users and verified TLS are preserved.

When Redis is bound to a private interface, enable an SSH tunnel. localhost in the host field then means the SSH server’s own loopback, which is how most self-hosted Redis instances are reached.

Where it stops

Redis Cluster and Sentinel profiles are not supported yet, and there is no key editor. For production debugging that is usually what you want. Read the guide on SCAN key browsing for details, or compare with the MongoDB connector.

Help articles

Frequently asked questions

Does DevOpsPad run KEYS * on my Redis server?

No. The key browser uses SCAN with a cursor and an optional MATCH pattern, and loads more keys on request. SCAN does not block the server the way KEYS does on a large keyspace.

Which Redis commands can I run?

Only an allowlist of read-only commands, including GET, MGET, GETRANGE, HGET, HLEN, LRANGE, LLEN, SCARD, ZCARD, XRANGE, HSCAN, SSCAN, ZSCAN, TYPE, TTL, PTTL, SCAN, DBSIZE, INFO and PING. Anything else is rejected before it is sent.

Why do I see the same key twice while scanning?

SCAN can return a key more than once, or return empty batches, while the keyspace changes. DevOpsPad keeps the cursor and removes duplicate names from the list it shows you.