Description:
There are a couple of features that I think if grouped together on within MEM would make our lives as DBA's much more easier in terms of Troubleshooting Database environments as well as providing meaningfull statistics from a Database Point of View to the Application developers for example.
How to repeat:
New Feature's
Suggested fix:
1. DB User Usage Statistics:
Usage by DB User’s (In Isolation and in Relation to total Instance Usage for All other Users)
• SQL Hits by User:
i. Total Transactions (highest to lowest with drilldown)
ii. Total Queries (highest to lowest with drilldown)
iii. Total Statements (highest to lowest with drilldown)
• User Stats per DB Schema, Per Table
i. Total User Connections
Frequency (How many times in a given period)
ii. Query Type (Drilldown for to SQL Hits by User)
Transactions
Queries
Statements
iii. Peaks & Lows
iv. Performance Degradation Contribution (%) – Per Query and for all queries that particular user runs over the entire DB Instance (Latency)
2. Data Carriage Statistics: (KB/MB/GB)
Total Data Transported within the DB @ the Instance Level: (Periods Apply of course)
• Total Data persisted into the:
i. InnoDB Buffer Pool
Doublewrite Buffer??
ii. Binary, Relay, Slow, General Logs
iii. Disk (Into data files)
Index
Data
• Total Data extracted out of the:
i. InnoDB Buffer Pool
ii. Disk (Out of data files)
Index (how Much of the Index file was Queried)
Data (how Much of the data file was Queried)
Data Temperature Monitoring (New Feature)
a. Cold, Cool, Warm & Hot Data Advisor (Based on above Index+Data Stats)
3. Storage IO Statistics on MySQL Devices: (Based on iostat tool)
• Total IOPS
• Wait %
• Wait Que Size
• Read Speed
• Write Speed
4. Slow Query Statistics:
• Slowest Query by:
i. Period
ii. Type
iii. User (Linked to Performance Degradation Contribution (%))
