Skip to content

Monitoring

  • Redis / Valkey

The monitoring tools sit in the Tools folder under a Redis connection: INFO dashboard, Slow log, Clients, Latency, Monitor and Big keys. In Cluster mode each tool can look at one node; the dashboard, MONITOR and big keys can also cover every primary.

The Redis INFO dashboard with live charts for operations per second, memory and clients, above the slow log listing the slowest recent commands with their durations.The Redis INFO dashboard with live charts for operations per second, memory and clients, above the slow log listing the slowest recent commands with their durations.
Watch the server live and find the commands that slow it down.

The dashboard polls INFO and keeps the history for the session as small charts.

Figure Shows
Memory Memory in use, of the limit if one is set, with peak, RSS and fragmentation.
Ops/sec Operations per second.
Clients Connected clients, and how many are blocked.
Hit ratio The recent hit ratio, and the ratio since the statistics were last reset.
Keys Keys in total, with a TTL, expired and evicted.
Network Traffic in and out, in KB/s.

Memory, Ops/sec, Clients and Hit ratio each have a chart of their history.

Below the figures, the keyspace table lists each Database with its Keys, keys With TTL and Average TTL, and Replication shows the server’s role (primary or replica), its primary and link status, or its connected replicas with their state, offset and lag.

  • Refresh every sets the polling interval: 1, 2, 5 (the default), 10, 30 or 60 seconds.
  • Pause stops polling; Resume continues.
  • Node picks one node in Cluster mode; by default the figures add up every primary.
  • Memory doctor shows the server’s MEMORY DOCTOR advice.

The toolbar also shows whether the server is Redis or Valkey, its version and its uptime.

  1. Open Tools › Slow log.
  2. Pick how many Entries to read: 32, 128, 512 or 1024. In Cluster mode, pick the Node.
  3. Read the ID, Time, Duration, Command and Client of each entry. Refresh reads the log again.

Reset… clears the slow log after a confirmation.

Tools › Clients lists the connected clients with their ID, Address, Name, User, DB, Age, Idle time, Last command and Flags. Client type narrows the list to Normal, Pub/Sub, Replicas or Master clients. Kill disconnects a client after a confirmation.

Tools › Latency shows the latency monitor’s events: for each Event, the Latest spike, the Latest latency and the All-time max, with a chart of its history.

  • The monitor records spikes above Threshold (ms, 0 = off). When it is 0, the panel says the latency monitor is off; enter a threshold and choose Set to turn it on.
  • Doctor shows the server’s latency advice.
  • Reset… clears the latency history after a confirmation.

Tools › Monitor streams every command the server runs. Choose Start MONITOR, then Stop MONITOR when you have seen enough; Clear empties the list. The panel keeps the latest 2,000 commands, and in Cluster mode shows which node ran each one.

Tools › Big keys samples keys with SCAN, sizes them with MEMORY USAGE and groups them by key pattern, with numbers and IDs replaced by *.

  1. Set the Sample size (5,000 by default) and, optionally, a key pattern.
  2. Choose Analyse. Cancel stops the run; the figures then cover the keys sampled so far.
  3. Read the summary: Sampled, Sampled memory, Estimated total and Time.

The Patterns table shows each pattern’s Keys, Memory, Share, Average, Largest and Types. Largest keys lists the 50 biggest keys with their Type, Memory, Length and TTL. If the server refuses MEMORY USAGE, the report says the sizes are unknown.

Big keys samples the live server. To see every key without touching the server, read an RDB file offline with Dump analysis. Before a run, the Big keys panel links to it: Analyse a dump file instead.

Documents Querybara 0.1.1 · built frombc9f5aa