OpenTelemetry in Focus, June 2023
Welcome back to OpenTelemetry in Focus for June, 2023! It’s officially summer, but it’s not just hot outside – we’ve had some major announcements and releases this month.
Are you a maintainer with something you’d like featured here? Get in touch with me via email, or on the CNCF Slack #otel-comms channel.
Releases and Updates
Here are the latest updates from some of our core repositories.
Specification
Version 1.22 includes a recommendation to reserve the aggregator normative for metrics, a move of the OTLP specification to the opentelemetry-proto repository, and an explanation of why custom attributes are not recommended to be placed in OTel namespaces.
There are no changes in other areas.
Collector
Version 0.80.0 has been released, along with 0.79.
0.80.0 includes a deprecation, enhancements, and module splits. The deprecation
is the service.PipelineConfig in favor of the pipelines.Config. Enhancements
include the addition of a dry run flag, allowing TLS settings to be provided in
memory, updates to the connector nodes, and updates to various modules.
0.79.0 includes deprecations, enhancements, and bugfixes. The release deprecates
the Host.GetExporters function, adds connectors to the output of the
components command, improves the behavior of the scraperhelper, optimizes the
multiBatcher to avoid a global lock, and fixes a bug related to data replication
in connectors.
In addition, collector-contrib has been released. Some notable changes include the addition of the WebSocket processor, updates to metric units in various receivers, and deprecation of options in the SumoLogic exporter. Check the release notes for more details.
0.79.0 included a number of bugfixes and enhancements, including breaking changes to hashing algorithms and changes to endpoint requirements. Enhancements include improved metric retrieval and new configuration options, while bug fixes address issues with subprocess cancellation and incorrect log stream filtering.
There were three releases of the Operator in June. The latest, release v0.80.0, introduces enhancements such as adding a Skywalking parser and populating credentials for Prometheus CR scrape configs. Bugfixes include a fix for the upgrade mechanism and empty global scrape interval, and a new component was added for NGINX auto-instrumentation in the operator.
0.79.0 of the Operator includes enhancements such as Prometheus metric exporter support for Node.js auto-instrumentation and the ability to inject the service version into the environment of the instrumented application. There is also a bug fix regarding the OpenTelemetry Collector version not displaying properly in the status field.
0.78.0 includes enhancements such as updating various packages, support for scaling on Pod custom metrics, and improved config validation. Bugfixes include addressing issues related to prometheus relabel configs and setting the default go auto-instrumentation version correctly. Components that were updated include the OpenTelemetry Collector, OpenTelemetry Contrib, Java auto-instrumentation, .NET auto-instrumentation, Node.js, Python, Go, and Apache HTTP Server.
Java
Version 1.27 has made the Log Bridge API and SDK stable. Important changes include merging log-related contents into the appropriate artifacts, marking the log SDK artifact as stable, and changing the default value of otel.logs.exporter in opentelemetry-sdk-extension-autoconfigure. The release also includes various bug fixes and improvements for the API, SDK, metrics, exporters, testing, SDK extensions, and semantic conventions.
Remember - the Log Bridge API is not intended for end users. It is used to bridge existing appenders (e.g., Log4j, SLf4J, Logback, JUL) into OpenTelemetry.
Java Instrumentation has been released, requiring some migrations to be performed.
- Migration notes: changes in Jersey 3.0 instrumentation suppression keys, renaming and splitting of opentelemetry-runtime-metrics artifact, deprecation of InetSocketAddressNetServerAttributesGetter and InetSocketAddressNetClientAttributesGetter, and introduction of new HTTP and network semantic conventions.
- New Java agent instrumentation for Quarkus RESTEasy Reactive and Reactor Kafka.
- Enhancements including improvements in Micrometer bridge, Ktor instrumentations, AWS SDK support, OkHttp 3, Jetty 11, Spring Boot, AWS Lambda tracing, and type matching.
- Bugfixes related to Logback MDC instrumentation, Kafka metrics reporter, Jetty context leak, filtering scalar Mono/Flux instances, and others.
JavaScript
Version 1.14
includes an optional forceFlush method to the SpanExporter interface.
.NET
Version 1.5.1 has been released, as well as 1.5.0. 1.5.1 is a fix for issues introduced in 1.5.0, and 1.5.0 consisted of the following major changes:
- The bug introduced in 1.5.0-rc.1 that caused the “Build” extension to return null has been fixed.
- The Exemplars feature has been marked as internal and will be added back in the 1.6.x prerelease versions.
- A new overload has been added for configuring MeterProviderBuilders while the IServiceCollection is modifiable.
- The Console and OpenTelemetryProtocol exporters no longer support exporting Exemplars.
- The ASP.NET Core instrumentation has been updated to fix an issue where baggage was cleared when the Activity stopped and to add a direct reference to System.Text.Encodings.Web.
- The HTTP instrumentation has been updated to fix an issue with missing metric data in network failures and to improve performance by avoiding boxing of common status code values.
Project and Community Updates
OTLP 1.0 is here
OTLP 1.0 is out (or will soon be)! We’ll publish more details about what this means on the blog at some later point. Keep in mind that just because the specification and protocol are 1.0, this does not mean all OpenTelemetry SIGs have adopted or released it. Please keep an eye out for future information.
This is still a major accomplishment, and a huge congratulations is in order to everyone who has contributed. Thank you!
From the blog
Here’s the latest blog posts. Want to contribute? Get in touch!
End-User Q&A: OTel at Farfetch is the latest in the End-User Q&A series, featuring Iris Dyrmishi of Farfetch.
Creating a Kubernetes Cluster with Runtime Observability discusses how to use recent tracing features in Kubernetes components to better understand your K8s clusters.
News and Upcoming Events
Observability Day is coming to KubeCon North America in Chicago! The CFP is open until August 6th. Submit your talk today!