This Discovery enhancement collects performance
metrics on probe/pattern and sensor processing times and then aggregates that data over time.
You can use the roll-up data to monitor the performance of specific discoveries or to compare
performance between versions after an upgrade.
Metrics
Discovery provides these individual performance metrics:
Probe and pattern processing time.
Sensor processing time.
Identification and Reconciliation Engine (IRE) processing time for Discovery patterns.
This processing time is already included in the sensor processing time, but is isolated
here to provided more insight into the identification and reconciliation of pattern
payloads.
Discovery can aggregate individual metrics for these attributes:
Builds/versions
Discovery status
Target IP address
How metric aggregations are triggered
Metric roll-ups are initiated as follows:
Aggregated by build: Implemented by the Aggregate
Discovery Probe And Sensor Metrics By Build scheduled job. This job runs
at 0200, local time.
Aggregated by status: Implemented by the Rollup
Probe/Sensor Metrics by Status Script Action, which is triggered by the
discovery.complete or
discovery.cancelled registered events.
Aggregated by target: Implemented by the Rollup
Probe/Sensor Metrics by Target Script Action that is triggered by the
discovery.device.complete registered event.
Note:
If Discovery execution is cancelled before it completes, Discovery cannot update
the IP target metric aggregation table. This is because the
discovery.device.complete event that triggers aggregation does not
run. IP target data for an interrupted Discovery is collected when subsequent discoveries
run successfully on the target. Cancelling Discovery execution does not affect the
aggregation of other metrics, which are triggered differently.
Tables
Discovery performance metrics data is stored in these tables:
Table
Description
Probe and Sensor Metrics (Individual)
[discovery_perf_metric_probe_sensor]
Stores the individual performance metrics for probes/patterns, sensors, and
IRE processing times.
Probe and Sensor Metrics (Aggregate)
[discovery_perf_metric_probe_sensor_rollup]
This is the base table for the metric aggregations and does not store data
itself. The three roll-up tables, by-build, by-status, and by-target all extend
this table.
Probe and Sensor Metrics (Aggregated by Build)
[discovery_perf_metric_probe_sensor_rollup_by_build]
Stores the aggregated performance metrics for probes/patterns, sensors, and
IRE by build and version.
Probe and Sensor Metrics (Aggregated by Status)
[discovery_perf_metric_probe_sensor_rollup_by_status]
Stores the aggregated performance metrics for probes/patterns, sensors, and
IRE by Discovery status.
Probe and Sensor Metrics (Aggregated by Target)
[discovery_perf_metric_probe_sensor_rollup_by_target]
Stores the aggregated performance metrics for probes/patterns, sensors, and
IRE by IP address.
Discovery properties
Performance metrics properties control whether or not aggregation occurs, but not what data
is included in the aggregation. Status and IP target data is collected as follows:
Rollups for status always contain new data. Discovery continuously collects data on
all probes and sensors during the discovery execution for that discovery status and
stores it in the Probe and Sensor Metrics (Individual)
[discovery_perf_metric_probe_sensor] table. Aggregation rolls up all probe and sensor
data for that particular status after discovery.cancel and
discovery.complete events are fired for that status, but only if
the aggregation property for status roll-ups is enabled.
Discovery continuously collects data on IP targets and stores it in the Probe and
Sensor Metrics (Individual) [discovery_perf_metric_probe_sensor] table. Aggregation
rolls up all existing IP target data after the
glide.discovery.perf.metrics.rollup_by_target property is enabled
and creates records in the Probe and Sensor Metrics (Aggregated by Target)
[discovery_perf_metric_probe_sensor_rollup_by_target] table.
These properties control the gathering of probe and sensor metrics:
Property
Description
glide.discovery.perf.metrics.enable_collection
Enables collection of performance metrics.
Type: true | false
Default: true
glide.discovery.perf.metrics.rollup_by_build
Enables aggregation of individual discovery performance metrics into a
roll-up table that groups by build.
Type: true | false
Default: true
glide.discovery.perf.metrics.rollup_by_status
Enables aggregation of individual discovery performance metrics into a
roll-up table that groups by discovery status.
Type: true | false
Default: false
glide.discovery.perf.metrics.rollup_by_target
Enables aggregation of individual discovery performance metrics into a
roll-up table that groups by target IP address. By default, Discovery continuously
collects individual IP address metrics, even when aggregation by target IP is
disabled. When you enable IP target aggregation, Discovery includes all target
metrics in the roll-up table.
Type: true | false
Default: false
View Discovery performance metrics for probes, sensors, and patterns
By default, Discovery tracks the
performance of individual probes, sensors, and patterns by measuring the processing
time. When patterns are used, Discovery measures the Identification and Reconciliation
Engine (IRE) processing time.
Before you begin
Role required: discovery_admin or admin
Procedure
Navigate to All > Discovery > Discovery Performance Metrics > Probe/Sensor (Individual).
Sort the list by Discovery status to see the list of probes and patterns that
ran in a specific Discovery.
You can view the metrics for each probe, sensor,
or pattern in the list or open the record.
All probe and sensor metrics data are
read-only.
The Probe and Sensor Metrics
(Individual) form provides these fields:
Field label
Field name
Description
Build/version
build_version
Build on which the Discovery was run.
Discovery status
discovery_status
ID number of the Discovery status from which the
metrics were collected.
Target IP address
target_ip
IP address of the target for this Discovery.
ECC queue input
ecc_queue_input
Identifies a particular ECC input record in the ECC
queue table.
ECC queue topic *
ecc_queue_topic
Identifies the Java class in the MID Server that
executes the probe.
ECC queue name *
ecc_queue_name
Identifies the probe/pattern evaluated for
performance in this aggregation.
Probe *
probe
Name of the probe used for this Discovery.
Probe processing time
probe_time
Interaction time of the MID Server with the target,
including construction of the payload that is sent back
to the instance. The time is in milliseconds.
IRE processing time
ire_time
Time required to process the pattern payload on the
instance by the Identification and Reconciliation Engine
(IRE). IRE time is useful because it shows the part of
the sensor time used by the pattern. The time is in
milliseconds.
Sensor processing time
sensor_time
Time it took the sensor to process the payload on the
instance for a Discovery. The time is in
milliseconds.
* Used to uniquely identify a probe/pattern and accompanying sensor
when gathering metrics for a Discovery.
View Discovery performance metrics aggregated by build
Use the roll-up by build data to ensure that the
processing times for Discovery components remain consistent for discoveries in a 24 hour
period. View aggregate build data before and after an upgrade to compare the performance
of the old and new versions. All aggregated performance data is
read-only.
Before you begin
Role required: discovery_admin or admin
Procedure
Navigate to All > Discovery > Discovery Performance Metrics > Probe/Sensor (Rollup-By-Build).
Sort the list by
Build/version.
Filter by a specific build to see the aggregated
processing times for the probes and patterns that performed a Discovery on
that build.
Open a record to see the probe/pattern statistics for
the selected build.
The form displays additional fields not visible on the
list. Roll-up calculations are over a 24 hour period, beginning every night
at 0200.
See the table of performance aggregation data for
descriptions of additional metrics displayed on the form for roll-ups by
build.
View Discovery performance metrics aggregated by status
Use the roll-up by status data to ensure that the
processing times for probes/patterns and sensors remain consistent for a specific
Discovery. All aggregated performance data is read-only.
Before you begin
Role required: discovery_admin or admin
About this task
Procedure
Navigate to All > Discovery > Discovery Performance Metrics > Probe/Sensor (Rollup-By-Status).
Sort the list by Discovery
status to see the aggregated processing times for the probes
and patterns that ran during a specific Discovery.
Filter by a specific status to display metrics
available for probes and patterns that ran in that status.
Open a record to see the probe/pattern statistics for
the selected status.
The form displays additional fields not visible
on the list. Roll-ups are only created for a completed or cancelled
status.
See the table of performance aggregation
data for descriptions of additional metrics displayed on the form for
roll-ups by status.
View Discovery performance metrics aggregated by IP address
Use the roll-up by target data to ensure that the
processing times for probes/patterns and sensors remain consistent for each Discovery of
a specific IP address. All aggregated performance data is read-only.
Before you begin
Role required: discovery_admin or admin
Procedure
Navigate to All > Discovery > Discovery Performance Metrics > Probe/Sensor (Rollup-By-Target).
Sort the list by Target IP
address.
Filter by a specific IP address to see the aggregated processing times for the
probes and patterns that performed the Discovery of that IP address.
Open a record to see the statistics about the selected
probe/pattern for the specific IP address.
The form displays additional fields not visible on the list. Roll-ups are
performed after IP Discovery has completed successfully.
See the table of performance aggregation data for
descriptions of additional metrics displayed on the form for roll-ups by
target.
Aggregated data for Discovery performance metrics
Discovery performance metrics can accumulate data for probes, patterns, and sensors
each time Discovery runs. Discovery calculates processing times and increments the number of
times a component runs for each roll-up profile: status, target, or build. All aggregated
performance data is read-only.
Sample roll-up form
This is an example of an aggregation record for probe and sensor metrics. The metrics
fields shown here are used for each aggregation.Figure 1. Sample roll-up by Status form
Performance Framework aggregated data
Except where noted, these fields are common to all aggregation records.
Field label
Field name
Description
Build/version
build_version
Name of the build running on the instance. This name identifies the version,
patch level, and release date of the ServiceNow platform.
Discovery status
discovery_status
ID number of the Discovery status record for this aggregation.
Note:
This
field only appears on the form for roll-ups by status.
Target IP address
target_ip_address
IP address of the target for this Discovery.
Note:
This field only appears on
the form for roll-ups by target.
Aggregation cutoff
aggregation_cutoff
The cutoff time varies, depending on the roll-up profile.
By-build: Occurs daily at 02:00, by default.
By-status: Closing time of the last aggregation for
that Discovery status, which might have run the last time
discovery.complete or
discovery.cancelled ran for that status.
By-target: Closing time of the last aggregation for
that target IP address, which might have run the last time
discovery.device.complete ran for that IP
address.
ECC queue topic *
ecc_queue_topic
Identifies the Java class in the MID Server that executes the probe.
ECC queue name *
ecc_queue_name
Identifies the probe/pattern evaluated for performance in this
aggregation.
Probe *
probe
Name of the probe used for this Discovery.
Probe time (count)
probe_time_count
Number of times a probe ran for a given roll-up profile.
Probe time (average)
probe_time_average
Average time a probe took to gather data on the target and format the payload
for a given roll-up profile.
Probe time (minimum)
probe_time_min
Minimum time a probe took to gather data on the target and format the payload
for a given roll-up profile.
Probe time (maximum)
probe_time_max
Maximum time a probe took to gather data on the target and format the payload
for a given roll-up profile.
Probe time (total)
probe_time_total
Total time used by a probe to gather data on the target and format the
payload for a given roll-up profile.
Sensor time (count)
sensor_time_count
Number of times a sensor processed payloads for a given roll-up
profile.
Sensor time (average)
sensor_time_average
Average time a sensor took to process payloads on the instance for a given
roll-up profile.
Sensor time (minimum)
sensor_time_min
Minimum time a sensor took to process a payload on the instance for a given
roll-up profile.
Sensor time (maximum)
sensor_time_max
Maximum time a sensor took to process a payload on the instance for a given
roll-up profile.
Sensor time (total)
sensor_time_total
Total time used by a sensor to process payloads on the instance for a given
roll-up profile.
IRE time (count)
ire_time_count
Number of times a pattern's payload was processed by the Identification and
Reconciliation Engine (IRE) for a given roll-up profile.
IRE time (average)
ire_time_average
Average time used for IRE processing of a pattern's payload for a given
roll-up profile.
IRE time (minimum)
ire_time_min
Minimum time used for IRE processing of a pattern's payload for a given
roll-up profile.
IRE time (maximum)
ire_time_max
Maximum time used for IRE processing of a pattern's payload for a given
roll-up profile.
IRE time (total)
ire_time_total
Total time used for IRE processing of a pattern's payload for a given roll-up
profile.
* Together, these values uniquely identify a probe/sensor pair (a "probe execution")
that is used for a Discovery.
