30 Jul 2019 08:13 AM - last edited on 18 May 2021 05:45 AM by MaciejNeumann
We are thinking about how to configure Dynatrace in a Black Friday scenario, I mean, it is expected that throughput will be much higher and in consecuence Response Time will be increased. Is it possible to configure Dynatrace to be compliance to this new scenario instead of followiwng its learned baselines? I am thinking on a global higher thresholds considering the new Peak Factor in the Anomaly Detection section.
If this is not possible, what the best practice to follow in a Black Friday or a promotional campaignwhen Peak Factor is expected to be x3 or even higher?
Thanks in advance.
Solved! Go to Solution.
In general I do not recommend tweaking anomaly detection just for one day. This may violate how AI will work day after as well. Another thing is that, if you have some configurations that are overwriting general settings on hosts / services level, then it may be hard to change such settings one way or another. We have such issues as well because of holiday, black fridays etc where traffic is changing from baselines. For now we have to deal with it.
I think best option would be recovering functionality from appmon, where we were able to prepare multiple configurations and change them via some schedule. In such option you can prepare configs for holiday / black friday etc. that will work for you. Good point for product idea. post one 🙂
We'd like to know how Customers handle this kind of scenarios. I'm sure, we are not the firts ones who raise this issue. If Dynatrace is not able to be configured in a Black Friday or vacatios Configuration settings, we'll be alert overloaded and Operations team cannot be notified for every Response Time baseline disruption because it is the expected behavior.
So, do you know any Customer who handles this scenarios?
Thanks in advance.
Our cistomers jest watch problem notifications and Puck those that are problematycznych like failure rates or huge response time degradations. Others are ignored.
But this is not handlable if problem notifications are integrated with Customer tools like Service Now, Jira, etc... it requires a person or a team looking Problems Dashboards, not working on a disregarded way....
I do not know automatic way to handle it. As I said, only idea I have right now is posting product idea about it (if there weren't any before).
Ok Sebastian. This is the RFE, please can you vote it?
I've already done it 🙂