How ISDA calculates Uptime for 24x7 services
(from an email 2/25/2010)
<!-- /* Font Definitions */ @font-face
@font-face
/* Style Definitions */ p.MsoNormal, li.MsoNormal, div.MsoNormal
.MsoChpDefault
@page Section1
div.Section1
-->Wikis: analyzed FY10 Q1 - Q3 for wikis production hosts (asterope/vega and confab), and counted anything that would result in unavailability from the user perspective (so either of database or web server, but not things like high swap utilization or NTP).
Q1:
Total: 7 hours, 41 minutes downtime (3 hours was during scheduled vega -> asterope cutover)
Assume 2160 hours in one quarter (30 * 3 * 24) = 99.6% uptime.
Q2:
Total 7 hours, 45 minutes (1 hour 15 minutes confab migration to OC11, 6.5 hours for asterope hardware issue)
Assume 2160 hours in one quarter (30 * 3 * 24) = 99.6% uptime.
Q3:
Total 1 hour = 99.9% uptime.
Stellar: application metrics for Nagios.
FY10 Q1:
7 hours downtime = 99.7% uptime.
FY10 Q2:
12 hours downtime = 99.4% uptime.
FY10 Q3:
7.5 hours total (4.5 for move to OC11) = 99.6% uptime
Data Warehouse:
FY10 Q1:
Total 44 hours = 98% uptime.
FY10 Q2:
3 hours downtime = 99.9% uptime.
FY10 Q3:
3 hours downtime = 99.9% uptime.