Questions
The questions graph shows the number of questions/minute. It has no direct relation to a tuneable parameter, but can be used as a hint to find out during which periods a more thorough logging should be conducted.
The most relevant types of questions (SELECT, INSERT, UPDATE and DELETE) are also logged.
Hide graphs
Connections
The connection graph shows the active connections and the configured max_connections.
In a sane configuration you should allways be able to see the max_connections graph. If it is hidden by the active connections graph all connections are in use.
Hide graphs
Tables
Table Cache
The table cache graph shows the open tables and the table_cache setting.
If the table_cache graph is hidden by the open_tables graph all table cache entries are in use.
Hide graphs
Temporary Tables
The temporary tables graph shows the number of implicit temporary tables. For the best performance these tables should be created in memory.
If there are many temporary disk tables (created_tmp_disk_tables), you can increase your tmp_table_size to improve your server's performance.
Hide graphs
Indexes
Key Cache
The key cache miss rate (key_read_requests/keyreads) should be lower than 1/100 (10m, 10 mili) indicated by an horizontal ruler and is quite ok for most systems if it's lower than 1/1000 (1m) also indicated by an horizontal ruler.
One could raise the key_buffer_size parameter to improve the key cache rate.
Hide graphs
Key Buffer
The key buffer graph shows the amount of memory that's in use as key cache.
Hide graphs
Full Joins
The select_full_join value should allways be 0. If its not 0, you should check the indexes of you tables.
The full joins graph can assist you in finding out when there have been full joins.
Hide graphs
Range Checks
The select_range_check value should allways be 0. If its not 0, you should check the indexes of you tables.
The range checks graph can assist you in finding out when there have been joins without keys where the key usage got checked after each row.
Hide graphs
Reads
The reads graph shows the read requests based on a key (handler_read_key) and the read requests based on a fixed position in a datafile (handler_read_rnd).
The handler_read_key indicates that your queries use properly indexed tables, this value should be high. The handler_read_rnd indicates the opposite and should be as low as possible.
Hide graphs
Slow Queries
The slow queries graph show the queries that took longer than long_query_time seconds (for this server the long_query_time=10 seconds).
If this value is not 0, you should activate the slow query log to find out which queries take that long to complete. The slow query log for your server is ON.
Hide graphs
Server Info
Your server (dbase1) has been running MySQL version 3.23.54-log for 425 hours since 01-03-2004 17:26.
General Remarks
Since startup, dbase1 received 37950460 questions.
| Question | Issued |
|---|---|
| select | 35580224 |
| change_db | 1460256 |
| admin_commands | 1347941 |
| show_status | 312114 |
| update | 272763 |
| insert | 215391 |
| show_variables | 86616 |
| show_tables | 10685 |
| show_keys | 3932 |
| show_fields | 2700 |
| show_create | 1670 |
| delete | 792 |
| show_databases | 110 |
| alter_table | 7 |
| create_table | 2 |
| drop_table | 2 |
| flush | 2 |
| set_option | 2 |
| analyze | 1 |
| check | 1 |
| grant | 1 |
| revoke | 1 |
| show_processlist | 1 |
Tuning Tips
Server tuning
Your server seems to be running for more than 7 days. It should be OK to use the tips below, but don't follow the advise blindly because this script is no replacement for common sense.
The maximum number of used connections is 257 which is 66 % of the configured maximum.
Your max_connections variable seems to be OK.
The table cache hit rate is 20 %, while 54 % of your table cache is in use.
Your table_cache variable seems to be OK.
The temporary table hit rate is 98 %.
Your tmp_table_size variable seems to be OK.
The key cache miss rate is 1/119271.
Your key_buffer_size seems to be too high, maybe you can use it's resources for something else.
Before you lower it, you should monitor the key buffer usage more closely. It may be possible that the keybuffer gets used at peak load.
Database tuning
The problems identified in this section can't be solved by tuning the server parameters. If you didn't design the database or wrote the queries yourself, you'll have to talk to the database developers to solve these.
- You need to check your indexes and queries because you had 3 full joins (select_full_join).
- You need to look at your indexes because you had 2144 joins without keys (select_range_check).
- You need to look at the indexing of your databases because the number of reads without using a key is too high. Reads using a key = 32209034 (handler_read_key), reads not using a key = 50407719 (handler_read_rnd).
- You've got 348 slow queries.
- Check your slow query log to find out which query is causing trouble.
