Intentionally dropping data from ActiveGate?

[mb996]

Hello DT Gurus,

We use AGs to route OA traffics to DT SaaS. Is it possible to configure the AGs to drop data, like dropping 100% of all data before reaching to DT SaaS? For those familiar with Splunk, this is similar to sending to nullqueue. 

We had a network outage and of course, the Network team "alleged" it was due to AGs egress to DT SaaS that saturated the network. 

Any suggestion is welcomed. 

Thank you.

[AntonPineiro]

Hi,

You can use network zones for adding traffic routing priorities.

Best regards

[PierreGutierrez · Accepted solution]

Hi @mb996 

I think that first it is necessary to analyze if many hosts are really being monitored.
Because one of the functions of the AG is to optimize the data packet sent by OA (reduce the weight), I think it is unlikely that AG is responsible.

I don't see much sense in deleting the data 100% before sending it to Dynatrace. If exactly that is needed, I understand that the monitoring of all Hosts should be turned off (mainly).

On the other hand, I share with you 2 alternatives to more efficiently control the observability of Requests.

1.- Settings/ Server-side service monitoring/URL-based sampling

   

With this option you can control the ATM impact on a specific endpoint and/or you can ignore the traffic of an endpoint.

Examples: 

2.- Settings/ Server-side service monitoring/ Deep monitoring / Exclude specific incoming web request URLs

 

I recommend you use the "deep monitoring" option and analyze which endpoints do not really generate value in your monitoring.  Mainly because the definition tells us that it is used to reduce bandwidth consumption.

Example :

Excluding request Health

Finally, I recommend creating a dashboard with ActiveGate monitoring metrics.
So you can have visibility of how the CPU, JVM, memory, incoming traffic, outgoing traffic, Dropped messages, forwarded messages , resent messages.

To have a more informed decision.

I hope you help 💪