Data gathering in HPE IMC 7.3

[ChristHP]

Hello,

IMC collects every 5mn, data of indexes for the monitored devices ( ex: CPU, RAM, ....)

When I display the data of the previous day, I got data every 5mn for the indexes.

When I display the date of 2 days before the current day, I got data every hour for the indexes.

I want to have data shown every 5mn for a week or a month in "Performance View". How can I manage it?

Thanks in advance for your help.

[jguse]

Hello,

IMC has a setting for this in the iMC\server\conf\qvdm.conf file (restart IMC after changing it):

# perf data keep time, its unit is day, and its range is: [2, 180].
# Warning: It is dangerous to change this value too big for disk.
perf_origindata_keep_time = 2

Please mind the warnings and carefully monitor the perf_db size after increasing it to ensure your DB server doesn't run out of free space. It can get very large even with the default setting of 2 days (20GB+ is not uncommon on larger installs), and aggregates the data over larger timespans by default to save space.

[ChristHP]

Hello Justin,

Thank you for your feedback.

Does it mean that If I want to keep the data every 5mn for 7 days, I have to set this parameter "perf_origindata_keep_time =7"?

Also, is it possible to reduce the data collection from 5mn to 30s for example?

Regards;

--

Christ

[jguse]

Does it mean that If I want to keep the data every 5mn for 7 days, I have to set this parameter "perf_origindata_keep_time =7"?

Yes, that is my understanding of how this feature should work. Documentation is very limited.

Also, is it possible to reduce the data collection from 5mn to 30s for example?

The polling interval is set to 5 minutes by default for every index. You can change this on a per-index basis via Performance Management > Global Index Settings > open the index settings and change the Interval to 30s (minimum).

Be careful, these kinds of changes rolled out to a lot of devices can quickly result in a huge perf_db and increased hardware requirements for IMC.