Hi Dynatrace Team,
In our environment we are implementing Dynatrace and we are exploring advanced features it has. Writing these questions to know more about this feature so we can implement this with best practice. From the Dynatrace product documentation I understand that Dynatrace One Agent has the capability to clean-up the files when the free disk space reaches 3% as part of emergency clean-up. The documentation doesn't have more details to understand this much better. Questions -
1. Does this cleanup take place as part of the "Low Disk Space" detection under Anomaly detection for infrastructure?
2. Is "Low Disk Space" detection customizable to a different value like 10%, 20% or 35% instead of 3% default?
3. Is it possible to ask Dynatrace to do this clean-up apart if the "Low Disk Space" detection is customized?
4. Files from what all Dynatrace filesystem directories will be cleaned-up as part of this feature, if enabled to monitor?
5. I see "Custom disk-detection rule" in Anomaly detection for infrastructure page. Is this option to monitor specific disks. Will this also do the cleanup for the free disk threshold I setup?
6. Now which one from the automatic detection and the custom disk detection takes precedence.
7. How to make sure that both the rules don't run on the server at the same time and create any adverse effects?
Any ideas / suggestions on how to test this feature on One Agent auto clean-up feature is much appreciated.
The link that explains the One Agent file ageing mechanism and log files clean-up is - https://www.dynatrace.com/support/help/setup-and-configuration/dynatrace-oneagent/oneagent-aging-mec...
@JDS After reading the article:
1 - This appears to run at 3% and not linked to the disk space alerting outside of Dynatrace cleaning up what it can.
2 - Low Disk space can be customized for an alert/notification perspective but has no bearing on the clean up mechanism for the Dynatrace OneAgent. If you want to change the clean up mechanism you will need to open a support ticket with Dynatrace Devs.
3 - That is exactly what it does. The clean up is set to run at 3% disk space left regardless to what your anomaly detection is set at.
4 - The support doc you linked provides all the directories that will be cleaned up. If you don't leverage something like log monitoring, that directory should be empty, therefore the mechanism wont have anything to clean up, but if it did, the clean up would apply.
5 - Custom Detection is set for you to define out specific rules per drive. This has no bearing on the clean up mechanism. The mechanism only applies to the drive in which the oneagent is installed on. If you installed in C, the clean up applies to the Dynatrace Directories in C. Not in D and not in folder trees outside of the oneagent. Simply put, Custom detection is for alerting only and has no connection to the clean up mechanism.
6 - They run congruently. If AI is set to automatically alert, it will trigger off the baseline. So say normal usage is 5% used. and you have a static threshold of 50% used alert. And the usage rises to 40%, while 40% is a deviation from the baseline, it will trigger an alert from the AI portion but wont trigger your custom event/metric. Now if usage jumps to 60% it is still a deviation and it hits/passes the static threshold so it will raise a second alert on the static setting you put. You don't need to have both, you can choose to have one or the other if needed.
7 - These Alerts don't run on the host, they run on Dynatrace SaaS so they wont impact the host.
This mechanism is coded with the OneAgent/AG and can only be changed with the help of the Dynatrace Dev Team. If you want to change the Mechanism you will need to open a support case with them. Just to be clear, Disk Anomaly detection is only for alerting purposes and has no affect on the Dynatrace Clean up Mechanism.