Remote Agentless WMI host limit

[mmevanson]

Greetings,

I noticed in the documentation for Agentless Windows Host monitoring that there is a remote host limit of 100!
That is way too low for us.   Is this 100 per active gate or 100 total?   
Is there a way to bump that limit up?   I assume it is there for resource considerations on the Active Gate.

Secondly - does anyone know what port this extension uses?   WMI port from AG (TCP 125)?

Thanks,

Matt

[rkdy · Accepted solution]

This would be per monitoring configuration; if you're monitoring, for example, 200 hosts, you can add 100 per monitoring configuration and have them run on the same ActiveGate group (just make sure your ActiveGate's in the Group have enough resources, otherwise split the load across separate activegate groups)

[gchastne]

Can a double-check this advice. We have previously been advised that this isn't about Extension Configurations and is a WMI issue - and hence a limit per ActiveGate.

[JamesKitson]

See the note on https://docs.dynatrace.com/docs/shortlink/extensions-concepts#limits-monitoring-configuration:

Remote WMI monitoring limits

While it's possible to enter up to 20,000 devices in a single configuration, remote WMI monitoring is limited by the number of simultaneous WMI queries that can be executed on the internal Windows service that handles them.

Currently, the limit is 100 queries. For example, if your extension uses 5 queries, configuring more than 20 devices may cause performance issues and gaps in the monitoring data.

We are working on ways to get around this Windows limitation in future versions of the WMI data source.

[gchastne]

Hi James,

Thanks for the details from the article. Unfortunately those are the words that lead to the question:

It doesn't define where the limit applies - at the configuration level, or at the ActiveGate level. 

The article starts with "single configuration", but then switches to "internal Windows service". If the issue is with the "internal Windows service" my assumption is that the issue is per ActiveGate because that's the level that the "internal Windows service" is running, but I don't know the specifics of the limitation.
Also, is the "metric query" defined within the Hub records for the Windows Remote Host monitoring the number we should be looking at, and hence 3 queries reduce the usage to 33 hosts. Remote Windows Host Monitoring monitoring & observability | Dynatrace Hub

[JamesKitson]

I don't work on the data source, but it's describing a limit that the Windows service can handle would be at the individual ActiveGate level. There is no 'framework limit' but the Windows service that is responsible for handling the extension activity can only handle a certain amount of activity so you can configure more than that but will likely see issues like gaps/missing data.

So the framework will let you configure many more than what WMI can support hence the note on the doc page.

And for the number of queries, this is the number of individual queries defined in the extension yaml (and enabled via feature set) and if all are enabled currently the supported extension has 15 queries defined which would mean based on this only about 7 would be supported per AG. It doesn't mean you will have gaps above this number but based on their testing of what the service can handle you may and this would be the cause.