newrelic / entity-definitions Goto Github PK

Summary

Opening enhancement request to add support for RF Scanner devices.
NOTE: # ( Provide a brief overview of what the new feature is all about. )

Desired Behavior

Possible Solution

Create RF Scanner entity. The required metrics are generated after receiving and parsing a scanner generated file in JSON format and ingesting via the Events API to create two custom event types.

Additional context

Description

The metrics in ext-host/golden-metrics.yml (e.g., https://github.com/newrelic/entity-definitions/blame/main/definitions/ext-host/golden_metrics.yml#L4) are malformed in a way that is not wrong but is likely misleading.

A metric has a single value; these have multiple values. . This is likely not what was intended.

The stream processing system that generates the golden metric data stored in NRDB will only use the first aggregation (in the example, it will construct newrelic.goldenmetrics.ext.host.systemLoad using only sum(datadog.system.load.1); the other values are ignored.

If those other values are important, then they should be included as separate metrics.

Steps to Reproduce

Expected Behavior

Relevant Logs / Console output

Your Environment

• ex: Browser name and version:
• ex: Operating System and version:

Additional context

Description

When navigating to Explorer, ElasticSearch entity type is not defined

Steps to Reproduce

1. Login to New relic
2. Click Explorer
3. On left navigation menu scroll down to On Host

Expected Behavior

Expected to see Elasticsearch
[NOTE]: # ( Tell us what you expected to happen. )

Relevant Logs / Console output

Metrics are currently captured by NR via on host infrastructure agent.

Your Environment

• Newrelic Infrastructure Agent 1.27.4

• nri-elasticsearch 4.5.3

• ex: Browser name and version:

• ex: Operating System and version:

Additional context

Definition is defined here: https://github.com/newrelic/entity-definitions/tree/main/definitions/infra-elasticsearchnode

Description

Only the Device Uptime is populating when viewing the UPS Entity dashboard for Vertiv UPS devices whose data is gathered via SNMP. Data for other fields is visible when you view the data via NRQL. Vertiv-ups.yml is a recent addition to the Kentik SNMP profile and likely needs to have a entity definition defined for it.

Steps to Reproduce

Search for and select a UPS entity to display the dashboard.

I am able to see all of the data for the Vertiv UPS's when I run a standalone NRQL query.
FROM Metric SELECT latest(kentik.snmp.PollingHealth), latest(kentik.snmp.lgpPwrNominalBatteryCapacity), latest(kentik.snmp.upsEstimatedMinutesRemaining), latest(kentik.snmp.upsBatteryTemperature), latest(kentik.snmp.gpPwrBatteryCapacity), latest(kentik.snmp.upsSecondsOnBattery), latest(kentik.snmp.upsBypassNumLines), latest(kentik.snmp.lgpPwrBatteryChargeStatus), latest(kentik.snmp.lgpPwrBatteryCharger), latest(kentik.snmp.upsInputLineBads), latest(kentik.snmp.upsOutputFrequency), latest(kentik.snmp.upsEstimatedChargeRemaining), latest(kentik.snmp.upsTestStartTime), latest(kentik.snmp.upsBatteryStatus), latest(kentik.snmp.lgpPwrBrownOutCount), latest(kentik.snmp.upsTestResultsSummary), latest(kentik.snmp.lgpPwrBatteryTestResult), latest(kentik.snmp.upsOutputSource), latest(kentik.snmp.upsTestResultsDetail), latest(kentik.snmp.lgpPwrBatteryLastCommissionTime), latest(kentik.snmp.upsAlarmsPresent), latest(kentik.snmp.upsBatteryVoltage), latest(kentik.snmp.lgpPwrBatteryTimeRemaining), latest(kentik.snmp.upsBypassFrequency), latest(kentik.snmp.lgpEnvTemperatureMeasurementDegF), latest(kentik.snmp.upsOutputVoltage), latest(kentik.snmp.upsOutputPower), latest(kentik.snmp.upsOutputPercentLoad), latest(kentik.snmp.upsInputCurrent), latest(kentik.snmp.upsInputFrequency), latest(kentik.snmp.upsBypassVoltage), latest(kentik.snmp.upsInputVoltage), latest(kentik.snmp.upsOutputCurrent), latest(kentik.snmp.lgpEnvTemperatureMeasurementDegC), latest(kentik.snmp.upsInputCurrent), latest(kentik.snmp.upsInputFrequency), latest(kentik.snmp.upsBypassVoltage), latest(kentik.snmp.upsInputVoltage), latest(kentik.snmp.upsOutputCurrent), latest(kentik.snmp.lgpEnvTemperatureMeasurementDegC), latest(kentik.snmp.lgpEnvTemperatureMeasurementDegF), latest(kentik.snmp.upsOutputVoltage), latest(kentik.snmp.upsOutputPower), latest(kentik.snmp.upsOutputPercentLoad), latest(kentik.snmp.upsInputFrequency), latest(kentik.snmp.Uptime), latest(kentik.snmp.upsOutputNumLines), latest(kentik.snmp.upsBatteryCurrent), latest(kentik.snmp.lgpPwrBlackOutCount), latest(kentik.snmp.upsInputNumLines) where provider='kentik-ups' FACET device_name since 5 minute ago LIMIT MAX

Expected Behavior

I would expect the built in Entity dashboard for UPS to be able to populate the Vertiv UPS data.

the formula should be:

latest(usec_per_read_op)/1000
latest(usec_per_write_op)/1000

rather than

latest(usec_per_read_op)/1024
latest(usec_per_write_op)/1024

ref: PR #14

Adding Unix host entity type based on newrelic-unix-monitor OHI.

Description

Using the dashboard validation script during local development (npm --prefix validator run sanitize-dashboards) is creating broken widgets on a specific query pattern seen in both the existing kentik_default dashboard and the new dashboard being added in PR #468

Steps to Reproduce

Clone the repo locally and run the dashboard validation script. No changes should be necessary as the current dashboard for Kentik Default entities is in "violation" and should show the problem.

Expected Behavior

This pattern should be ignored by the validation script since it's a valid dashboard definition and NRQL syntax before being adjusted.

Relevant Logs / Console output

This is only happening on the following pattern:

"FROM Metric SELECT round(latest(kentik.snmp.hrStorageUsed)*100/latest(kentik.snmp.hrStorageSize), .01) AS 'Used %', round(latest(kentik.snmp.hrStorageSize)*1e-6,.01) AS 'Total (MB)', round(latest(kentik.snmp.hrStorageUsed)*1e-6,.01) AS 'Used (MB)' WHERE provider = 'kentik-poweriq' FACET storage_description LIMIT MAX"

Specifically; this portion of the string:

(kentik.snmp.hrStorageSize), .01)

Which is being improperly sanitized to this: (replacing ) with } and removing ,)

(kentik.snmp.hrStorageSize} .01)

Additional context

This is negatively impacting each time a PR is run and the sanitization script is executed locally as it puts these dashboards at risk of a breaking change in production with a bad NRQL syntax on these widgets.

I am first timer. I need more information or examples on how I can tie NRQL alert condition to a synthesized entity.
Customer using custom events/attributes in the NRQL alerts.

Summary

I'm not sure how useful golden metrics such as average(cpuPercent) are generally, so I'd like to propose some new ones.

• Cpu time spent waiting for I/O (while non-idle): percentile((cpuIOWaitPercent / cpuPercent) * 100, 50, 75, 95, 99, 99.9)
• CPU time spent waiting for VM (wile non-idle) percentile((cpuStealPercent / cpuPercent) * 100, 50, 75, 95, 99, 99.9)

Also, what is the policy about percentiles with golden metrics? I don't see many using percentiles. Is a golden metric of percentile(cpuPercent, 50, 75, 95, 99, 99.9) an option?

Desired Behavior

Averages for golden metrics aren't terribly useful in my opinion. Would like to see more percentile based golden metrics so I can compare the median to outliers.

Possible Solution

If a percentile() with multiple percentiles specified is discouraged, could do one percentile per golden metric.

Additional context

Can submit a PR, just looking for some guidance on the best approach.

Description

Summary metrics for Unix host appear empty (CPU, memory, network)

Steps to Reproduce

Solaris hosts are reporting metrics, but this entity dashboard is empty save for the entity names/account info:

The backing metrics are present, as unixMonitor:Process, unixMonitor:Stats, etc, and the host entities exist.

Expected Behavior

Expected summary metrics to appear in the metric columns on the dashboard.

Your Environment

• Solaris

Additional context

Metrics queries here: https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-unix_host/golden_metrics.yml

Query for Memory relies on memory.used, but these hosts aren't sending it. A workaround could be to use other data, but I'm not sure if this is only an issue for a particular OS. For example this query works instead on the same data, by calculating used memory as total minus free:

FROM `unixMonitor:Stats` select ((sum(memory.total) - sum(memory.free)) / sum(memory.total)) * 100 as 'Memory utilization (%)' facet hostname timeseries

Since NR Infra (and OHI) agents doesn't support AIX platform, Customers use FIT InfraDBPluginJava (https://source.datanerd.us/FIT/InfraDBPluginJava) to monitor databases like DB2, Oracle and other DBs. These entities are not visible in the Entity Explorer or Workloads.

I would like to create Entity Synthesis Definition for DB2 and Oracle database for Manulife.

I have customer using Prometheus heavily for Redis instances in their Kubernetes cluster.

I tried creating the config file for this specific case, but I tried to stay away from Kubernetes specifics so I expect this to also work for Redis outside of Kubernetes.

ref: #5

Summary

I want to add one or more entity synthesis definitions to support the AWS service of database migration service (DMS) so that we can see DMS instances, DMS tasks, and DMS tables as entities.

Possible Solution

Create 3 different entities

1. DMS Instance - Track each instance which holds many tasks which run many tables. There are many metrics like CPU and rows per second which are meaningful to track at a higher level than a DMS task.
2. DMS Replication Task - Track each task for replication which has 1 source endpoint and 1 target endpoint for 1 or more tables. Would want to track the name, task stats, logs, etc. here.
3. DMS Table - Track each table and its counts and status.

Additional context

Description

The following definitions summary metrics are rendering as null:

https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_compellent/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_recoverpoint/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_vmax/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_vnx/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_vplex/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-dell_xtremio/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-pure/summary_metrics.yml
https://github.com/newrelic/entity-definitions/blob/main/definitions/ext-synology/summary_metrics.yml

Steps to Reproduce

Navigate to the Explorer view, select an entity grouping, look at the summary page:

Expected Behavior

Summary metrics populated per the definitions.

Relevant Logs / Console output

Your Environment

• ex: Browser name and version:
• ex: Operating System and version:

Additional context

Ref: PR #31

Adding rcluster host metrics

Summary

Hello, I want to add an entity definition Mikrotik routers (https://mikrotik.com), but have doubts on the best way of doing it.

Desired Behavior

Either have a specific Mikrotik router entity type, or have them in a generic router type.

Possible Solution

I see that there currently is an EXT-ROUTER definition, which code talks about Kentik routers, but also has this comment

• # Leaving this ambiguous to allow future technical partners to use the same entity type with additional conditions

Should I:

1. submit a new entity definition,
2. or modify the EXT-ROUTER to synthesize also the data from the Mikrotik routers?

If answer is 2:

• How do I add a mikrotik condition to the type definition?
• How do I modify these summary metrics to pick the metrics starting with mikrotik.? For example, this is what we have for Kentik:

 - cpuUtilization:
  title: CPU Utilization (%)
  unit: PERCENTAGE
  query:
    select: average(kentik.snmp.CPU)
    from: Metric
    where: "provider = 'kentik-router'"
    eventId: entity.guid

• How do I modify the dashboards definition to condiitonally use metric names kentik.x.y or mikrotik.a.b? FOr example:

"nrqlQueries": [
  {
    "accountId": 0,
      "query": "FROM Metric SELECT average(kentik.snmp.CPU) FACET entity.name WHERE provider = 'kentik-router' TIMESERIES 5 MINUTES"
  }
],

Additional context

The router's data is dumped via the Metrics API with a script that runs on the router itself (pretty much like an agent), found here https://github.com/OscarDCorbalan/mikrotik-newrelic.

By impersonating userId 2848364, you should see a dashboard Mikrotik Router in the EU region (one.eu.newrelic) that uses that data.

This the shape of the data that is sent:

[
   {
      "common":{
         "attributes":{
            "mikrotik.currentfirmware":"6.48.2",
            "mikrotik.model":"RBD52G-5HacD2HnD",
            "mikrotik.serialnumber":"C6140C7EBEF8",
            "mikrotik.upgradefirmware":"6.48.2"
         },
         "timestamp":1620969329
      },
      "metrics":[
         {
            "name":"mikrotik.firewall.connection.established",
            "type":"gauge",
            "value":32
         },
         {
            "name":"mikrotik.firewall.connection.tcp",
            "type":"gauge",
            "value":32
         },
         {
            "name":"mikrotik.firewall.connection.udp",
            "type":"gauge",
            "value":11
         },
         {
            "name":"mikrotik.interface.ether1.fprxbps",
            "type":"gauge",
            "value":656
         },
         {
            "name":"mikrotik.interface.ether1.fptxbps",
            "type":"gauge",
            "value":2736
         },
         [ ... rest of interfaces' throughput data ...] 
         {
            "name":"mikrotik.ip.dhcpserver.leases",
            "type":"gauge",
            "value":3
         },
         {
            "name":"mikrotik.ip.dns.cache.size",
            "type":"gauge",
            "value":1024
         },
         {
            "name":"mikrotik.ip.dns.cache.used",
            "type":"gauge",
            "value":589
         },
         {
            "name":"mikrotik.ip.pool.used",
            "type":"gauge",
            "value":2
         },
         {
            "name":"mikrotik.system.cpu.load",
            "type":"gauge",
            "value":2
         },
         {
            "name":"mikrotik.system.memory.free",
            "type":"gauge",
            "value":80629760
         },
         {
            "name":"mikrotik.system.memory.total",
            "type":"gauge",
            "value":134217728
         }
      ]
   }
]