Having enough.
WHAT TO MONITOR
Processor use Memory use Disk space Disk activity Network throughput Connection counts
WHAT TO ALERT ON
Disk approaching full Memory exhaustion Sustained high load
WHY DISK FIRST
A full disk stops everything, corrupts things, and is entirely predictable.
WHAT FILLS DISKS
Logs Backups kept locally Uploads Temporary files Database growth
WHAT TO CONFIGURE
Log rotation Cleanup of temporary files Retention limits on backups
WHAT TO PROJECT
Growth, from observed trends.
WHAT TO PLAN FOR
Predictable peaks Gradual growth
WHAT TO DO BEFORE ADDING RESOURCES
Establish what is consuming them.
WHY
Inefficiency is frequently cheaper to fix than to accommodate.
WHAT TO REVIEW
Resource use monthly, against capacity.
WHAT TO RECORD
The trend, so growth is visible.
