Server Status view — MySQL

A health read on the server, from its own status counters.

MySQL exposes several hundred global status counters, named as if you already know what you're looking for. The Server Status view pulls the ones that answer "how is this server doing?" into a simple summary on the left, and groups the rest so the full set is still browsable.

Server Information #

This section identifies the server, which helps when you're about to change something and need to be sure which machine you're on.

Field	What it shows
Version	Version and build description, which tells MySQL from MariaDB or Percona.
Host, Port, Socket	The server's host name, and the port and socket it listens on.
Basedir, Datadir	Where the installation and data files live. Useful for finding configuration and log files.
SSL	Whether the server has SSL support. See SSL Options.

The Summary #

Field	Counter	Reading it
Running Since	`Uptime`	When the server last started. Every other counter is measured from here.
Status Since	`Uptime_since_flush_status`	When the counters were last reset with Flush Status.
Bytes Sent, Bytes Received	`Bytes_sent`, `Bytes_received`	The total amount of data sent and received. Most useful as a rate.
Threads Connected	`Threads_connected`	The number of client connections open right now.
Peak Connections	`Max_used_connections`	The highest number of connections open at once since the server started. Compare with Max Allowed.
Max Allowed	`max_connections`	The configured maximum number of connections. If Peak Connections reached it, connections have been refused.
Connection Attempts	`Connections`	The number of connection attempts since startup, successful or not.
Failed Connections	`Aborted_connects`	The number of attempts that never completed: wrong credentials or host, handshake timeouts.
Dropped Connections	`Aborted_clients`	The number of established connections lost without closing cleanly. A steady increase usually means clients timing out or crashing.
Threads Running	`Threads_running`	The number of connections actively executing. The real measure of load: 400 connections with 3 running is idle.
Threads Created	`Threads_created`	The number of connections that needed a brand-new thread because no cached one was free. If this is close to Connection Attempts, almost every connection is paying that cost, and a larger thread cache would help.
Threads Cached	`Threads_cached`	The number of idle threads kept ready for reuse.

Flush Status #

The Flush Status button runs FLUSH STATUS, resetting most counters to give you a fresh window to measure. Status Since gives you the timestamp of the last time the stats were updated.

All Status Variables #

Next to the summary view is a list of every global status variable grouped by name prefix. Ungrouped variables may appear in the Miscellaneous group.

Recipes #

To tell whether the server is under load

Read Threads Running, not Threads Connected.

Hundreds of connections with a handful running is a pool at rest.
If Threads Running is persistently high, see what they're doing in Processes.
Check the Commands group for the mix of work.

A Com_select-heavy server needs different attention than a Com_insert-heavy one.

To find out whether connections are being refused

Compare Peak Connections with Max Allowed. Equal means the ceiling was hit.
Check Failed and Dropped Connections.

Failures point at credentials or networking; drops point at clients going away mid-session.
Raise max_connections in Variables only after confirming the connections are legitimate.

A leaking pool fills any ceiling.

To size the thread cache

Compare Threads Created with Connection Attempts.

If Threads Created is a large fraction of Connection Attempts, most connections get a new thread instead of a cached one.
Raise thread_cache_size in Variables.
Flush Status, let the workload run, and check the ratio again.

To measure a specific period

Flush Status right before the period begins.
Let it run. Status Since confirms the window.
Read the counters. They now describe just that period.