Primo Uptime Report for Primo MT AP01 Instance (APAC) - Q2 2018
Uptime Quarterly Reports are published each quarter to provide a comprehensive view of our uptime performance as measured over the last three (3) months and over the last twelve (12) months. The report measures our performance as defined in the Service Level Agreement.
This document details the uptime report for the Primo MT APAC01 in APAC.
The uptime measured on a rolling 12 month basis July 2017 - June 2018 is 99.90%
The uptime measured on a rolling 3 month basis April 2018 - June 2018 is 99.79%
Please note our System Uptime Status page, which allows our customers to view the current status of their instance/environment at any time, using the following link: http://status.exlibrisgroup.com
Unscheduled downtime incidents in Q2 2018
|
Date |
Start time [SGT] |
End time [SGT] |
Duration |
Description |
|---|---|---|---|---|
| April 03, 2018 | 01:40 PM | 02:14 PM | 34 |
An extreme rare issue in the application caused high load on the search engines which led to high memory usage. |
| April 06, 2018 | 11:22 AM | 12:10 PM | 48 | |
| April 09, 2018 | 12:00 PM | 13:00 PM | 60 |
A high load on the oracle data base lead to a sever service interruption. The root cause to the high load on the oracle was a results of a frequent and un-optimized application SQL query and the increased size of the oracle table that lead to a change in the query execution plan. After the first event we enabled advanced Oracle monitoring which highlight the problematic query. Ex Libris DBA add index to the table that prevent full table scan and reduce the load on the oracle. |
| April 16, 2018 | 13:06 PM | 14:02 PM | 56 | |
| April 30, 2018 | 09:43 AM | 10:47 AM | 64 | As part of the investigation of April 30th disrupton, Ex Libris developers took memory dump on May 1st that caused the service disruption on that day. Following deep investigation, it was found that a software malfunction in the area of timeout management lead to a chain reaction, where when a single search node was slow the others hang. |
| June 24, 2018 | 17:10 PM | 18:22 PM | 72 |
The reactive database rules initiated a new execution plan to use a full table scan which affected the search functionality in a way that made the system unavailable To solve this the DB was restarted to restore the previous plan. |
Scheduled downtimes during maintenance windows in Q2 2018
|
Start Date |
Start Time [SGT] |
End Time [SGT] |
Duration (Minutes) |
|---|---|---|---|
| April 08, 2018 | 00:12 AM | 00:23 AM | 11 |
| April 22, 2018 | 00:00 AM | 01:00 AM | 60 |
| May 13, 2018 | 00:20 AM | 04:00 AM | 220 |
| May 13, 2018 | 04:35 AM | 06:43 AM | 128 |
| May 13, 2018 | 22:00 PM | 03:00 AM | 300 |
Total unscheduled downtime minutes during past 12 months
|
Quarter |
Total unscheduled downtime in Quarter (minutes) |
|---|---|
|
Q3 2017 |
140 |
|
Q4 2017 |
0 |
|
Q1 2018 |
54 |
|
Q2 2018 |
630 |
How is Uptime Calculated?
The uptime calculation is based on the following calculation (as defined in the Service Level Agreement):
"Uptime" means the total period in minutes during which the Service is available for access and use during this period.
“Uptime Percentage” means Uptime expressed as a percentage, calculated in accordance with the following formula:
Uptime Percentage = X /(Y–Z) × 100
Where:
X = Uptime
Y = Last 12 months period
Z = The duration (in minutes) of any SLA Exclusions* during these 12 months
*SLA exclusions are defined in the contract SLA (e.g. Scheduled maintenance, etc.)
Further Information
If you have any queries on the information within this report please contact Ex Libris through your usual channel.