Build 175

2020-04-21

On-Premises Release Notes

Performance Improvement

Span meta-data is now written directly to file, previously we have stored it in Cassandra.

Cumulative Features from Release #174, #175

Applications, services and endpoints dashboards

The "Infrastructure Issues & Changes" chart replaced the "Top Traces" widget on the dashboard "Summary" tab. The new chart can help you find correlations between infrastructure issues or changes and application metric changes, e.g. increase of "Erroneous Call Rate" or "Latency". For more information, see the Infrastructure Issues & Changes docs.

Infrastructure Issues & Changes chart

The "Latency Distribution" Chart now gives you the possibility to jump directly to Unbounded Analytics with filters already selected for the timebucket you click.

Clickable Latency Distribution Chart

Native AWS Lambda Tracing for Python

In addition to the existing Node.js support, Instana now offers automatic tracing of AWS Lambda functions built in Python. Instana's native tracing of AWS Lambda functions provides large advantages over X-Ray based tracing:

  • The execution of Lambda functions seamlessly correlates with tracing of other monitored components: Calls from Hosts processes to Serverless and back are now linked, correlated and analyzed as such with no additional steps required.
  • Instana provides rich performance analysis between specific versions of the same AWS Lambda function. Compare and contrast performance of various versions of your Lambda functions.
  • Reduces Cost/Lower AWS Bills: Completely eliminates the need of pulling costly X-Ray data from AWS.
  • Provides automatic and deep trace visibility into the activities of your Lambda functions. A big improvement over the limited and very manual X-Ray alternative.

New built-in metric for Kubernetes Deployment and OpenShift DeploymentConfig

We added a new built-in metric for Deployment and DeploymentConfig, Available to Desired Percentage, that allows you to create a custom event based on a percentage value of the ratio of available to desired replicas if the existing built-in event Available replicas is less than desired replicas for Deployment or DeploymentConfig does not meet your needs.

Kafka Connect

Instana now supports monitoring of Kafka Connect, meaning that you can monitor performance of your worker instances, deployed connectors, and sink as well as source tasks running. This includes a built-in dashboard that displays key metrics about your instances' availability, connector and task startup failure and more. For more information, see the Kafka Connect docs.

Vault Sensor

Instana now supports monitoring of HashiCorp Vault. This includes a built-in dashboard that displays key metrics about your instances’ availability, secrets, tokens, audit logs, engine errors and more. For more information, see the Vault docs.

VMware Tanzu, Cloud Foundry, Pivotal Container Service (PKS)

  • Tracing Ruby applications in Cloud Foundry is now generally available.
  • Tracing PHP applications in Cloud Foundry is now experimental; the tracing itself works very well and is entirely automated; we need to perform a few changes to correlate the infrastructure correctly.

Improvements

  • The "Process Top List" on the host overview page shows the percentages aligned to the number of CPUs.
  • The event detail page of CPU usage events shows the "Process Top List" if a host is involved.
  • OpenShift 4.1, 4.2 and 4.3 support.
  • Our website monitoring JavaScript agent now ignores Angular specific user timings by default.
  • The services and endpoints lists now offer a button to jump to Unbounded Analytics with preset filters (type, technologies, and free-text search).
  • Jumping from Log Messages to Unbounded Analytics now keeps around the free-text search.
  • A detailed error is shown on Unbounded Analytics when grouping on a tag is not supported.
  • The VMware Tanzu foundation name is now visible in the Cloud Foundry dashboards and can be filtered for.
  • PKS clusters are marked as Managed by Pivotal Container Service in the Kubernetes dashboards.
  • The origin of the Instana agents in the Agents view in Instana is now tile; the Instana agents installed on PKS worker nodes will be reported as public docker for consistency with other Kubernetes deployment means.
  • Kafka sensor: Improved auto-discovery of server.properties file path and enabled specification via sensor configuration.
  • RabbitMQ sensor: Enhanced support for selecting which queues are going to be monitored with regular expressions.
  • Added Kubernetes 1.17 support.

Fixes

  • Website Smart Alerts: Fixed infinite loading indication of the chart in case there is no underlying data.
  • PHP-FPM sensor 1.1.59: add support for bracketed ipv v6 addresses
  • PHP-FPM sensor 1.1.59: prevent connections to unreachable FPMs
  • PHP-FPM sensor 1.1.59: add support for --fpm-config and other long arguments