Cloud integrations release notesRSS

New

• New Relic now collects labels for Google Compute Engine, Kubernetes Engine, Cloud SQL, and Cloud Functions services. They join BigQuery, Cloud Spanner, Cloud Pub/Sub, and Cloud Storage integrations, which were collecting labels already. With the appropriate permissions, you'll be able to use the following metadata to narrow down queries, dashboards, and alerts related to Google Cloud resources:
  • Labels for Functions.
  • Labels for SQL databases.
  • Labels for Kubernetes clusters, which are propagated to nodes, pods and containers. If you'd like entity specific labels to be collected, you can use the Kubernetes integration.
  • Labels and tags for Compute Engine virtual machines, and labels for disks. Tags are treated as a regular label with value true.
• Additionally, Inventory for Google Compute Engine and Google Cloud Functions has been enhanced with new attributes and restructured, so it's easier to find the metadata that belongs to each entity.
• The aforementioned improvements might cause an unexpected volume of Entity modified events, as Inventory will be updated with the new data.

Bug fixes

• Some Google Kubernetes Engine entities had duplicate identifiers. In order to fix this issue, New Relic has regenerated the internal entityId and externalKey attributes for all containers in Insights events and Infrastructure inventory for this cloud integration.

  • This change doesn't have any impact either on the integration default dashboard provided by New Relic Infrastructure or on the Insights queries.
  • However, all entities related to Google Kubernetes Engine will be created again in Inventory. Accordingly, you might see duplicated entities for 48 hours, until the old identifiers expire. This might also cause an unexpected volume of Entity created events.

• Metadata for Google Compute Engine virtual machines was not being added to the metric events reported by the New Relic infrastructure agent. In order to fix this issue, New Relic has regenerated the internal externalKey attribute for all virtual machines and disks in Insights events and Infrastructure inventory for this cloud integration.

  • This change doesn't have any impact either on the integration default dashboard provided by New Relic Infrastructure or on the Insights queries.
  • However, all entities related to Google Compute Engine will be created again in Inventory. Accordingly you might see duplicated entities for 48 hours, until the old identifiers expire. This might also cause an unexpected volume of Entity created events.

• For Google Cloud integrations, the value of the zone attribute was not reported consistently. As a consequence, some inventory attributes and event metadata, such as project, were not reported for some Google Cloud Storage buckets.

New

The account status dashboard for cloud-linked accounts introduces some significant changes:

• The Account changes chart lists any action that was performed on the linked cloud account configuration, from linking, renaming and enabling cloud services for monitoring, to modifying the polling intervals or any other data collection option.
• The charts in the dashboard have been reorganized to better understand API call usage, errors that affect New Relic collecting data, and changes to the integrations configuration.
• For Google Cloud-linked projects, the dashboard now includes the Stackdriver calls chart, which shows the amount of API calls that New Relic sends to the Google Stackdriver Monitoring API.

New

• In order to optimize the volume of calls to AWS APIs, the cloud integrations setting options have been enhanced. In particular:
  • New Relic will be using the same polling interval for metrics and inventory data from now on.
  • Many options for you to control the amount of inventory data that New Relic collects from your cloud account have been added: tags, extended inventory, regions, entity name prefixes, resource groups, etc. Please find which attributes are fetched by enabling extended inventory in the specific document for each cloud integration.
• You can enable or disable the collection of extended inventory attributes through New Relic's user interface for AWS DynamoDB, AWS ElastiCache and Google Cloud BigQuery.
• The AWS Lambda integration has been extended with additional metrics and inventory attributes:
  • The ServerlessSample event type now includes deadLetterErrors and iteratorAge metrics, both for functions and function aliases. Additionally, it includes the concurrentExecutions metric for functions that have a custom concurrency limit specified.
  • For functions, inventory (under aws/lambda/function/) now includes the kmsKeyArn, masterArn, revisionId and layers attributes.
  • For function aliases, inventory (under aws/lambda/alias/) now includes the revisionId and routingConfig attributes.
  • For mappings between an AWS resource and a function, inventory (under aws/lambda/event-source-mapping/) now includes the stateTransitionReason attribute.

Bug fixes

• Google BigQuery tables with no schema were lacking some inventory attributes.

New

• Google Cloud Load Balancing integration is now available. New Relic collects metrics for all types of load balancers, and provides a curated dashboard with successful requests, error percentages, throughput, and latency. Check Google Cloud Load Balancing monitoring integration for details.
• Google Cloud Pub/Sub integration is now available. New Relic collects metrics and inventory data for Pub/Sub Topics and Subscriptions. Check Google Cloud Pub/Sub monitoring integration for details.
• Google Cloud Spanner integration is now available. New Relic’s curated dashboard shows relevant metrics for Spanner instances and databases, such as request latency, disk, CPU, successful requests count and error rate. Check Google Cloud Spanner monitoring integration for details.
• Google Cloud SQL is now available. The most relevant metrics to watch are CPU, memory and disk utilization, to make sure the databases are correctly dimensioned. The curated dashboard also features write and read operations per second, concurrent connections and database state. Check Google Cloud SQL monitoring integration for details.
• Google BigQuery is now available. Our curated dashboard helps keep track of query execution time, the number of bytes stored in the datasets, and the number of bytes uploaded to any table in the dataset that were billed. Check Google BigQuery monitoring integration for details.

New

• The AWS EMR integration has been optimized to reduce the number of API calls that are made to collect data for terminated clusters, while ensuring that the latest metrics and inventory attribute values will be reported.

• The AWS Route 53 integration has been optimized to reduce the number of calls to the AWS API that are needed to fetch hosted zone data. It's just one call per 100 hosted zones now, while an additional call per hosted zone was made before.

• In the AWS RDS integration, the inventory attribute allocatedStorageBytes has been added to the RDS instance metric events, so now it's possible to calculate free storage percentage with an NRDB query:

  SELECT min(provider.freeStorageSpaceBytes.Minimum * 100 / provider.allocatedStorageBytes) FROM DatastoreSample where provider = 'RdsDbInstance' FACET displayName TIMESERIES AUTO

  Copia

Bug fixes

• When creating an Integrations alert condition through the user interface, some data sources were wrong or missing in the list:
  • For the Google Kubernetes Engine integration, GcpKubernetesPod and GcpKubernetesNode data sources should be used instead of GcpKubernetesPodSample and GcpKubernetesNodeSample, respectively.
  • For the Google Cloud Pub/Sub integration, GcpPubSubSubscription and GcpPubSubTopic data sources should be used instead of GcpPubSubSubscriptionSample and GcpPubSubTopicSample, respectively.
  • For the Google Cloud Functions integration, GcpCloudFunction data source should be used instead of GcpCloudFunctionSample.
  • For the AWS Lambda integration used together with AWS CloudFront, the LambdaEdgeFunction data source was missing.
• In the AWS VPC integration, the value of the provider attribute for PrivateNetworkSample events, was fixed from VpcEnpoint to VpcEndpoint.
• For the AWS API Gateway integration, the Count metric aggregate, which provides the number of calls to API methods, has been changed from Sum to Sample Count. This change requires that you manually update all the alert conditions and custom Insights dashboards that involve the deprecated Sum metric.
  • To update an alert condition, use New Relic Infrastructure Alerts UI or API and edit the metric that defines the threshold for the ApiGatewayApi data source. If you use the UI, please replace Provider - Count - Sum with Provider - Count - Sample Count. If you use the API, please replace provider.count.Sum with provider.count.SampleCount.
  • To update an Insights dashboard, use New Relic Insights UI or API and edit the chart queries that involve the provider.count.Sum to replace it with provider.count.SampleCount.
  • In particular, if you enabled AWS API Gateway monitoring before December 2018, New Relic might have created a "AWS API Gateway (APIs)" dashboard automatically. This dashboard contains the "Total Calls compare with 1 hour ago" chart, which involves the provider.count.Sum metric that needs to be replaced.
  • Please note that the default dashboards shown under New Relic Infrastructure have already been updated.

New: Enable or disable attribute collection via UI

You can enable or disable the collection of extended inventory attributes through New Relic's user interface for AWS Application Load Balancer (ALB), Elastic Load Balancer (ELB), Elastic Beanstalk, Elastic Block Store (EBS), Route53 and Simple Notification Service (SNS). This tool helps you reduce the amount of calls that New Relic does to AWS APIs, and it might contribute to contain your AWS bill. For more information, see the documentation for:

• Configuring the polling frequency of data collection through the infrastructure integrations UI
• Attributes that make part of extended inventory in each cloud integration

New: View errors fetching data via UI

Also, the cloud integrations Account status dashboard shows any errors New Relic might experience when trying to fetch inventory and metric data for your cloud resources. The Fetching errors chart now shows the cloud service these resources belong to. This will help to facilitate troubleshooting the integration setup.

Changes in curated dashboards

New Relic infrastructure integrations will not automatically create new dashboards in New Relic Insights when an integration is enabled.

Instead, curated dashboards for On-host and Cloud integrations data will be embedded in the New Relic Infrastructure UI, and they can be reached from the following links:

• On-host integrations:
  https://infrastructure.newrelic.com/accounts/<your_account_ID>/integrations/onHostIntegrations
• AWS integrations:
  https://infrastructure.newrelic.com/accounts/<your_account_ID>/integrations/aws
• Azure integrations:
  https://infrastructure.newrelic.com/accounts/<your_account_ID>/integrations/azure
• GCP integrations:
  https://infrastructure.newrelic.com/accounts/<your_account_ID>/integrations/gcp

Please refer to the transition guide and the Infrastructure integration dashboards and charts documentation for more details about the new pre-built integration dashboards.

Note the existing curated dashboards that had been automatically created in New Relic Insights through the InfrastructurePro@newrelic.com user are still available. These dashboards won't be automatically updated anymore, but now you can edit and remove them. By default, they have the same name as the dashboard in New Relic Infrastructure, that takes this format:

• For on-host integrations: <Integration name>
• For cloud integrations: <Integration name> - <Linked account name>

Other changes

• The former systemErrors metric in the AWS DynamoDB integration is now reported in several different metrics which represent the total number of requests that generate an HTTP 500 status per operation type. Please refer to AWS DynamoDB monitoring integration for details.

New

• The AWS RDS integration now provides new metrics for RDS database instances and Amazon Aurora clusters. Check AWS RDS monitoring integration for details.
• The Azure App Service integration now provides new metrics to monitor connections and performance of Web Apps. Check Azure App Service monitoring integration for details.

New

• It's now possible to filter by tags in the Infrastructure default dashboards. In the filter, tag keys and values are grouped under the Labels category so they can be more easily found.
• The AWS ALB integration now collects tags for ALB Target Groups that can be used in NQRL queries and dashboard filters. You can also fine-tune the data gathered with this integration by specifying the resource tag key and value you want to monitor with the new Filter by tag. Please refer to Configure polling frequency and data collection for cloud integrations for details.
• The AWS API Gateway integration now collects tags for stages, resources and methods, which can be used in NQRL queries and dashboard filters.
• The AWS Redis Cache integration now collects the OperationsPerSecond metric, which reports the number of commands processed per second by the cache server.

Changes

• The threshold selector in Infrastructure Integrations Alerts now includes the metric unit, so it's much easier to configure the correct threshold value. It is available for RDS metrics, and shall be available in other services soon.
• The BucketSizeBytes metric for AWS S3 buckets was returning only the size of Standard Storage objects contained in the bucket. Now the metric returns the aggregated size of all storage types. Note the BucketSizeBytes metric is included in the DatastoreSample event type with a provider value of S3Bucket.