| The name of the server shown above is Brubeck. MTBF calculated these statistics on 4/10/98 at 5:00:11 a.m., as shown in the "Sample was taken" field. Brubeck runs build 4.1.5
The Start Date column shows you how far back in time MTBF went when compiling the statistics. You might request statistics for the last year, but if MTBF only logged information for the last few months, then this column shows that date. The End Date column shows the most current log entry date that MTBF found when calculating statistics for the period requested. Brubeck has been running for a little over a year.
The other columns include information on the number of shutdowns and crashes, and the percentage of time the server was up and down, between the start date and the end date. You can see that Brubeck had 17 crashes, and 34 shutdowns during this time. There were 48 startups. The percentage of time Brubeck was up and running was 93.46%.
You can also look down the rows and find other statistics on Brubeck:
Range Covering: This row shows a breakdown of the time between the Start Date and the End Date.
Average Time Up Between Failures: This row shows a breakdown of the average time the server was up and running between the Start Date and the End Date. This number may not be a good way to measure server performance, because quickly restarting your servers after all crashes could make this number deceptively high.
Geometric Mean Between Failures: This row shows a breakdown of the geometric mean of the time the server was up and running between the Start Date and the End Date. For example, if we had 11 measurements sorted by the least amount of time up to the most amount of time up, the geometric mean would be the uptime of the sixth measurement, since it falls right in the middle.
Maximum Time Up: This row shows a breakdown of the maximum time the server was up and running between the Start Date and the End Date.
Minimum Time Up: This row shows a breakdown of the minimum time the server was up and running between the Start Date and the End Date (including server shutdowns).
Average Time Up: This row shows a breakdown of the average time the server was up and running between the Start Date and the End Date (including server shutdowns).
Total Time Up: This row shows a breakdown of the total amount of time the server was up and running between the Start Date and the End Date (including server shutdowns).
Maximum Time Down: This row shows a breakdown of the maximum time the server was down between the Start Date and the End Date (including server shutdowns).
Minimum Time Down: This row shows a breakdown of the minimum time the server was down between the Start Date and the End Date (including server shutdowns).
Average Time Down: This row shows a breakdown of the average time the server was down between the Start Date and the End Date (including server shutdowns).
Total Time Down: This row shows a breakdown of the total amount of time the server was down between the Start Date and the End Date (including server shutdowns).
Server Last Up: This row shows a breakdown of the amount of time the server was running when MTBF calculated these statistics.
Viewing adjusted server statistics
The second section of charts in a Server Statistic document contains adjusted server statistics. These charts have the same rows and columns as the set of server statistics charts we just examined. The only difference is that if a Server Crash document is marked resolved, MTBF does not factor that crash into these statistics. For example, if there is a power failure and your server was not on a UPS, MTBF generates a Server Crash document. You could mark the crash as resolved, since it was not Notes-related and it was explainable.
At Iris, we use this section of charts after we fix a bug that caused multiple server crashes. We then mark all the Server Crash documents for those crashes as resolved. The next time MTBF generates statistics, it no longer takes the resolved crashes into consideration when calculating the adjusted server statistics. We can then get a good idea of our average uptime with the bug fixed. This lets us know how close a build is to a release. The following charts show adjusted statistics: |