Could someone please help explain where the code goes for Extension 2.0?
I've played with OneAgent and ActiveGate extensions by writing my own Python code that runs every 1 minute. I understand a fair amount about Extensions 2.0 YAML, JSON, and ZIP thanks to this great blog post (https://www.dynatrace.com/news/blog/extend-dynatrace-automation-and-ai-capabilities-more-easily-than...) and help site (https://www.dynatrace.com/support/help/extend-dynatrace/extensions20/ ).
But, I can't seem to figure out where I place the logic that knows what to do with the YAML and JSON configuration data to obtain metrics or perform some logic on data to produce a metric. Is maybe Extensions 2.0 not a replacement for what we can currently do with ActiveGate and OneAgent extensions?
Solved! Go to Solution.
Thank you for the reply. Looking forward to the coded extensions. We're playing with the SNMP extension right now and it's really cool. The best part is whoever developed it I feel was really thinking about all the different scenarios a client might have and thus made the SNMP extension quite flexible. I can't image the complexity of code behind it.
My experience has been to enable SNMP extension 2.0 and also create an extension for Linux host based on the structure I had of the SNMP yaml, with host related metrics.
Both extensions are up and running, after a rather complicated signing process for custom extension. This is an issue to improve or simplify as it is very cumbersome.
My question is about the promise that extension 2.0 provides "metrics with topology". When I read topology I immediately relate it to entities displayed in the smartscape. But so far I do not see any results in that sense or a Host monitored with SNMP in the smartscape.
In the structure of the extension at the end comes the Topology and Relationships section where I have that the network interfaces are children of the SNMP device (Host). Could you explain to me what would be the scope of these metrics with topology. I imagine a scenario where I monitor a Switch, and establish a relationship between a switch interface and a Host. If that interface throws errors, it could generate a Problem that would impact that Host.
I had the same thought as you when i saw Topology.
I too am ultimately interested in the same. First just so I can relate switches to routers and such so that the network topology is clear. But yes, the best state would be to link the switch interface to a host.
Hopefully DT will come back with a response.
From what I have seen in some of the example from Dynatrace, there is both a topology, relationship and other keys in the xmls.
The thing is that the documentation is extremely lacking at the moment.
It seems it should also be possible to query other wmiNamespaces, but so far I have not been able to do this, and there is no logging which provides any clues.