NAM monitor metrics

Metrics and dimensions

Dimensions (all network availability monitors)

Monitor metrics (all network availability monitors)

Step metrics (all network availability monitors)

Request metrics (all network availability monitors)

ICMP monitor metrics

TCP monitor dimensions

TCP monitor metrics

DNS monitor dimensions

DNS monitor metrics

DQL queries to extract data

Latest Dynatrace

Use DQL queries to extract data with metrics and dimensions.

The following examples use Notebooks to demonstrate how to work with the results of network availability monitors. They use result metrics as a starting point to explore possibilities of extracting and interpreting data with DQL.

Host entity status for ICMP requests

The monitor MULTIPROTOCOL_MONITOR-5C2F92334DF71A90 executes ICMP requests and filters monitored hosts with "targetFilter": "hostGroup == e2e-synthetic-private-location" (which resolves to about 26 hosts).

By using the dt.synthetic.multi_protocol.request.executions metric and splitting it by the dt.entity.host and result.status.message dimensions, we can display the status of the connection to a particular monitored host entity in this host group. Some hosts do not fulfill the expected success rate; instead of being HEALTHY, their requests are marked as CONSTRAINT_VIOLATED.

Timeseries status = avg(dt.synthetic.multi_protocol.request.executions),
by:{dt.entity.host, result.status.message},
filter: dt.entity.multiprotocol_monitor == "MULTIPROTOCOL_MONITOR-5C2F92334DF71A90"

Number of ICMP packets sent and received

The monitor MULTIPROTOCOL_MONITOR-548C3CD54183CED9 executes ICMP requests to hosts with the explicitly defined IP addresses, 18.x.x.x, 10.x.x.x, and 34.x.x.x. Each of these IP addresses maps to a distinct host.

We use the sum of the dt.synthetic.multi_protocol.icmp.packets_sent and the sum of the dt.synthetic.multi_protocol.icmp.packets_received metrics to get insight into how many packets were sent and received.

We split results by the request.target_address dimension and filter data for 18.x.x.x and 10.x.x.x only.

For 18.x.x.x, the same number of packets were received as were sent, but for 10.x.x.x, all packets are lost and none were received.

timeseries {
  packets_sent = sum(dt.synthetic.multi_protocol.icmp.packets_sent),
  packets_received= sum(dt.synthetic.multi_protocol.icmp.packets_received)
},
by:{request.target_address},
filter: dt.entity.multiprotocol_monitor == "MULTIPROTOCOL_MONITOR-548C3CD54183CED9"
    AND (
        request.target_address == "18.x.x.x"
        OR request.target_address == "10.x.x.x"
    )

Target status for TCP requests

The monitor MULTIPROTOCOL_MONITOR-74E68F22FF5E9227 executes TCP requests to hosts from a host group that resolves to the IP addresses 18.x.x.x, 34.x.x.x, and 44.x.x.x.

To display the status of the TCP connection for an IP-port pair, we use the dt.synthetic.multi_protocol.request.executions metric and splitting it by the dimensions:

In this example:

Note that results of this query can be limited to just successful requests by filtering by the multi_protocol.result.status.code dimension (code == 0).

timeseries status = sum(dt.synthetic.multi_protocol.request.executions),
by: {request.target_address, request.tcp_port_number, result.status.message},
filter: dt.entity.multiprotocol_monitor == "MULTIPROTOCOL_MONITOR-74E68F22FF5E9227"
//  and result.status.code == 0

TCP connection time to target port

The monitor MULTIPROTOCOL_MONITOR-74E68F22FF5E9227 executes TCP requests to hosts from a host group that resolves to the IP addresses 18.x.x.x, 34.x.x.x, and 44.x.x.x.

In this example, instead of IP addresses, we split the results by monitored host entity IDs.

To check typical time taken for a successful connection to the target port for a host, we use the average of the dt.synthetic.multi_protocol.tcp.connection_time metric split by the dimensions:

Only ports 22 (SSH) and 8080 (HTTP server) are open, and these are the only ports for which dt.synthetic.multi_protocol.tcp.connection_time is available. The hosts are effectively in different geographical locations (Ohio, Oregon, and North Virginia in the United States), so a difference in connection times is expected.

timeseries duration = avg(dt.synthetic.multi_protocol.tcp.connection_time),
by:{dt.entity.host, request.target_address, request.tcp_port_number},
filter: dt.entity.multiprotocol_monitor == "MULTIPROTOCOL_MONITOR-74E68F22FF5E9227"

Execution statuses

All network availability monitors

TCP monitor execution statuses