Performance Metrics Scratchpad

From Siwiki

Contents 1 Performance Metrics 2 Scratch Pad 2.1 Things we want to observe 2.2 Suggestions for Metrics 2.2.1 Service Metrics 2.2.2 Resource Metrics 2.3 Metric Pages

[edit] Performance Metrics

This is a collection of standard metrics for operating system performance observability.

This is a draft page for a draft concept. Revisit this page in n number of weeks to check on progress (where n is a positive integer).

[edit] Scratch Pad

This is a scratch pad for ideas on the performance metric collection.

[edit] Things we want to observe

latency.

[edit] Suggestions for Metrics

are the "prstat -mL" process microstates suitable to become standard metrics?

Need cool pics/icons for performance degradation profiles: gradual, steep, negative exponential, ... Draw using inkscape and convert to png using "rsvg".

[edit] Service Metrics

• latency / response / service time
• queue length
• throughput
• errors

eg, NFS, LDAP, DNS, web

[edit] Resource Metrics

• utilization
• saturation
• errors

eg, CPU, Disk, Net

[edit] Metric Pages

Performance Metric 00001: CPU Utilization by System