05 Sep 2014 07:50 AM
Hello,
Not sure if this is the right forum but would appreciate some help. We have Dynatrace 5.5 and I'm trying to change the minimum disk space thresholds from the default 3% and 1GB for Host monitors. Found this link but still having problems getting the new values to work:
I've placed my settings in dtserver.ini, and restarted but still only getting the default thresholds:
1) Do i have to specify the units of measurement in the parameters? example if i wanted minFree space less than 10GB: -Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreshold=10 ? or 10GB?
2) What happens if I only want to set the threshold for MinFreePercent and ignore space? How is that written?
3) How do i modify this default disk alert to send more than just the default ONE notification? I want it to send multiple times until monitor is good again.
4) Am I supposed to use the misspelled Threashold or correctly spelled Threshold spelling? I am using Dynatrace 5.5
Thanks!
09 Sep 2014 02:02 PM
moved to dynaTrace Forum
09 Sep 2014 05:42 PM
Hi,
For memory, I had to use bytes. For example: -Dcom.dynatrace.diagnostics.healthMemAvailableThreshold=500000000
Then restart the dynaTrace server.
As for free space, why not set it to a very low value such as 1 Megabyte?
For the misspell, it seems it will work with both in future versions but might be safe to stick to the misspell for now as Version 6.0 Host Health Monitoring says even 6.0 GA had the misspell.
Josh
25 Oct 2014 07:21 PM
Hi,
I 'm trying to change the same parameter but Dynatrace is just accepting GB units and I need to set MB units, example:
Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreshold=3000000000 (Ok for 3 GB)
Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreshold=5000000000 (Ok for 5 GB)
Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreshold=500000000 (not Ok for 500MB)
Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreshold=2500000000 (not Ok for 2,5 GB)
Juan
27 Oct 2014 06:15 PM
Hi,
I confronted R&D with it. Bear with me for the one or other hour.
One thing I could confirm: The somewhat awkward unit Byte stems from the fact that it was only meant for dev. With other settings you can specify other units like MB or GB.
Thanks
G.
04 Nov 2014 03:16 PM
jcurbina@mainsoft.cl if you put any characters other than decimals at the end of this parameter parsing will fail and the value is defaulted to 1GB. You have to specify the value in bytes.
04 Nov 2014 03:22 PM
The shown tooltip in the UI is a rounded value therefore 500000000 will be shown as 0GB!
Regards,
Georg
04 Nov 2014 11:58 PM
Thanks for your information. I could confirm also with support that we must use bytes units. Also the Alert will be raised when both condition is met (% free space and free space Mbytes).
Juan
05 Nov 2014 12:28 PM
Hola, Juan!
I just updated the documentation with the info I gleaned from this thread and after verification with the dev.
To summarize everything in this thread I put the text right below.
We are aware of the clunky UI, but it will not yet change in the upcoming release!
Thanks a ton for pointing this out and for your patience!
I´d kindly ask you for feedback on the new text!
G.
To change the behavior, define different values for the thresholds in <DT_HOME>/dtserver.ini
. E.g.
to set the memory-percentage-available-threshold to lower-than-default 5% and / or
-Dcom.dynatrace.diagnostics.healthMemPercentageAvailableThreshold=5
to set the memory-available-threshold to lower-than-default 500000000 Bytes (roughly 500MB or .5 GB).
-Dcom.dynatrace.diagnostics.healthMemAvailableThreshold=500000000
Absolute sizes must be stated in Bytes, not MB or GB.
All sizes must not have any units attached. Settings other than purely numeric will be ignored.
If you want to ignore one of two ANDed values, just set an unreachable value.
In the example above: 0 for the percentage or 0 for memory size.
05 Nov 2014 05:53 PM
Hi Guenter,
Thanks a lo tfor your support.
I'm a little bit confused because I have the following response about a support ticket where they told me is an AND between both conditions, so setting healthDiskMinFreePercentageThreashold=0, I believe is not going to work.
"This is working as designed.
Alert will be raised when both condition is met.
1) -Dcom.dynatrace.diagnostics.healthDiskMinFreePercentageThreashold=1
2) -Dcom.dynatrace.diagnostics.healthDiskMinFreeSpaceThreashold=500MB
You need to change healthDiskMinFreePercentageThreashold value
if you wants to trigger alar when disk space is less than 500MB.
To trigger disk alarm with 1%, disk size needs to be larger than 50GB"
In my case our customer have several filesystem with less than 30 GB and they want to be alerted when disk free space is 500 MB.
Thanks,
Juan
06 Nov 2014 04:09 PM
Hi Juan,
thanks for the feedback that made us actually take a pencil and paper and make a truth table to find out we had it upside down. <fg>
What did we (here at Dynatrace) learn?
Question dev statements, use your own brains and TRY things (at least on paper, but then you rely on dev statements again).
Hope that helps
G.
To change the behavior, define different values for the thresholds in <DT_HOME>/dtserver.ini
. E.g.
to set the memory-percentage-available-threshold to 5%. (higher than default
-Dcom.dynatrace.diagnostics.healthMemPercentageAvailableThreshold=5)
-Dcom.dynatrace.diagnostics.healthMemAvailableThreshold=500000000
to set the memory-available-threshold to 500,000,000 Bytes (roughly 500MB or .5 GB; lower than default).
Absolute sizes must be stated in Bytes, not MB or GB.
All sizes must not have any units attached. Settings other than long integer will be ignored.
There are conditions like Alert when memory available < n% AND < nBytes
. If you want to ignore one of two ANDed values, just set a threshold that will be reached for sure. In the example above: Set a very high percentage or size. Memory available < 100%
will always return true, so lastly an alert will be triggered when memory available < nBytes
becomes true.
Please mind in testing: The whole memory group is considered unhealthy when either memory available is too low or page faults per second are too high. So an alert could also stem from page faults and not from memory available.
07 Nov 2014 01:01 AM
Hi Guenter,
Great!
This is working as the client wanted, It could be perfect if the alert message includes the filesystem name who trigger the incident. I suppose is needed an enhancement for this.
Thanks a lot for your help.
Best Regards,
Juan Urbina
17 Mar 2015 10:06 AM
Heyz ,
Is there away to remove one of the values completely ? like if I want to check only based on disk % or if there away to update the alert caption that says "at least one disk less than .. and less than ..
Thanks
23 Mar 2015 01:07 PM
Saad,
I´m not sure if I understand exactly.
Q: "If I want to check only based on disk %..."
A above: "If you want to ignore one of two ANDed values..."
Please elaborate, if I misunderstood!
G.