Hello. There are a couple of things wrong which I will try to explain.
First, let's understand the Service entity. It does not actually have an attribute called "ipAddress" as we do not register IP Addresses against services. Going by relation, the closest entity that holds an IP Address is the Host entity. This means that you can "walk" the relationship from the Service to the Host (if the database has a monitored host). If your database only has a Service, extracting an IP Address is not possible.
Second, some of the DQL syntax is incorrect. The `dt.entity.` prefix only applies to entity types which ipAddress is not, so we cannot use `dt.entity.ipAddress`.
To extract the IP address from the Host, we'll first get the Host IDs and then lookup their IP Addresses.
timeseries interval: 1h,
ResponseTime = avg(dt.service.request.response_time),
by: { ServiceId = dt.entity.service },
filter: {
// Adapt the selector to look for Services related to Hosts
in(dt.entity.service, classicEntitySelector( """type("SERVICE"),databaseVendor("DB2"),fromRelationships.runsOnHost(type("HOST"))""" ))
}
// Add the service name (for readability) + Host IDs
| fieldsAdd ServiceName = entityName(ServiceId, type:"dt.entity.service")
// Service may run on multiple hosts so we need to 'expand' the array
| expand host = entityAttr(ServiceId, "runs_on", type:"dt.entity.service")[dt.entity.host]
// Now we can lookup each host's IP Addresses
| fieldsAdd IpAddresses = entityAttr(host, "ipAddress", type:"dt.entity.host")
// Remove unnecessary fields and keep result like this to see IPs per service instance
| fields timeframe, interval, ServiceId, ResponseTime, ServiceName, IpAddresses
// OR aggregate it to see IPs across all service instances
| summarize {
ResponseTime = avg(ResponseTime[]),
IpAddresses = arrayFlatten(collectArray(IpAddresses))
},
by: { timeframe, interval, ServiceId, ServiceName}
I hope this helps you out.
