Skip to content

v0.123.0
Compare
Choose a tag to compare
@srv-gh-o11y-gdi srv-gh-o11y-gdi released this 17 Apr 18:06
· 111 commits to main since this release
v0.123.0
This tag was signed with the committer’s verified signature.
dmitryax Dmitry Anoshin
GPG key ID: 5F41515F29C651CC
Verified
Learn about vigilant mode.
a93426c
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

This Splunk OpenTelemetry Collector release includes changes from the opentelemetry-collector v0.123.0
and the opentelemetry-collector-contrib v0.123.0
releases where appropriate.

❗ Known Issues ❗

🛑 Breaking changes 🛑

  • (Splunk) Support for OpenSUSE 12 and CentOS 7 (tar package) has been dropped (#6073)
  • (Core) service/telemetry: Mark telemetry.disableAddressFieldForInternalTelemetry as beta, usage of deprecated service::telemetry::address are ignored (#12756)
    IMPORTANT: If you have service::telemetry::address configured manually, ensure to change it to the new format: 
    telemetry:
  metrics:
    readers:
      - pull:
          exporter:
            prometheus:
              host: <HOST>
              port: <PORT>
              without_scope_info: true
              without_type_suffix: true
              without_units: true
    Otherwise, the Collector will not emit any internal telemetry metrics.
    To restore the previous behavior disable telemetry.disableAddressFieldForInternalTelemetry feature gate.
  • (Core) exporterbatch: Remove deprecated fields min_size_items and max_size_items from batch config. (#12684)
  • (Contrib) receiver/azuremonitor: multi subscriptions support and automatic discovery (#36612)
  • (Contrib) exporter/kafka: change default client_id to "otel-collector" (#38411)
    The exporter now uses the "configkafka" package which consolidates
    common configuration structures and default values. As a result of
    this change, we update the default client_id value to "otel-collector".
  • (Contrib) receiver/kafkametrics: change default client_id to "otel-collector", deprecate "refresh_frequency" (#38411)
    The receiver now uses the "configkafka" package which consolidates
    common configuration structures and default values. As a result of
    this change, we update the default client_id value to "otel-collector",
    and deprecate "refresh_frequency" in favour of "metadata.refresh_interval".
  • (Contrib) processor/k8sattributes: Remove stable feature gate k8sattr.rfc3339 (#38810)

🚩 Deprecations 🚩

  • (Core) exporters: Deprecate blocking in favor of block_on_overflow. (#12710)
  • (Core) exporters: Deprecate configuring exporter batching separately. Use sending_queue::batch instead. (#12772)
    Moving the batching configuration to sending_queue::batch requires setting sending_queue::sizer to items
    which means that sending_queue::queue_size needs to be also increased by the average batch size number (roughly
    x5000 for the default batching configuration).
    See https://github.com/open-telemetry/opentelemetry-collector/tree/main/exporter/exporterhelper#configuration
  • (Contrib) exporter/kafka: Deprecate auth.plain_text configuration. Use auth.sasl with mechanism set to PLAIN instead. (#38883)
  • (Contrib) receiver/kafkametrics: Deprecate auth.plain_text configuration. Use auth.sasl with mechanism set to PLAIN instead. (#38883)
  • (Contrib) receiver/kafka: Deprecate auth.plain_text configuration. Use auth.sasl with mechanism set to PLAIN instead. (#38883)

💡 Enhancements 💡

  • (Splunk) Add discovery rule (disabled by default) for istio (#5854)
  • (Core) exporters: Add support to configure batching in the sending queue. (#12746)
  • (Core) exporters: Add support for wait_for_result, remove disabled_queue (#12742)
  • (Core) exporters: Allow exporter memory queue to use different type of sizers. (#12708)
  • (Core) service: Add telemetry.newPipelineTelemetry feature gate to inject component-identifying attributes in internal telemetry (#12217)
    With the feature gate enabled, all internal telemetry (metrics/traces/logs) will include some of
    the following instrumentation scope attributes:
    • otelcol.component.kind
    • otelcol.component.id
    • otelcol.pipeline.id
    • otelcol.signal
    • otelcol.signal.output
      These attributes are defined in the Pipeline Component Telemetry RFC,
      and identify the component instance from which the telemetry originates.
      They are added automatically without changes to component code.
      These attributes were already included in internal logs as regular log attributes, starting from
      v0.120.0. For consistency with other signals, they have been switched to scope attributes (with
      the exception of logs emitted to standard output), and are now enabled by the feature gate.
      Please make sure that the exporter / backend endpoint you use has support for instrumentation
      scope attributes before using this feature. If the internal telemetry is exported to another
      Collector, a transform processor could be used to turn them into other kinds of attributes if
      necessary.
  • (Core) exporters: Enable support to do batching using bytes sizer (#12751)
  • (Core) service: Add config key to set metric views used for internal telemetry (#10769)
    The service::telemetry::metrics::views config key can now be used to explicitly set the list of
    metric views used for internal telemetry, mirroring meter_provider::views in the SDK config.
    This can be used to disable specific internal metrics, among other uses.
    This key will cause an error if used alongside other features which would normally implicitly create views, such as:
    • not setting service::telemetry::metrics::level to detailed;
    • enabling the telemetry.disableHighCardinalityMetrics feature flag.
  • (Contrib) connector/spanmetrics: Add instrumentation scope to span metrics connector. (#23662)
    This change adds the instrumentation scope to the span metrics connector, which allows users to specify the instrumentation scope for the connector.
    Now, the connector has a new configuration option:
    • include_instrumentation_scope: A list of instrumentation scope names to include from the traces.
      The instrumentation scope name is the name of the instrumentation library that collected the span.
  • (Contrib) receiver/splunkhec: Splunk HEC timestamp accepts nanoseconds, microseconds, milliseconds, and seconds epoch. (#36571)
  • (Contrib) receiver/oracledb: Add support for additional read/write metrics (#37814)
    The following metrics are now available, all disabled by default:
    'physical reads direct', 'physical writes', 'physical writes direct',
    'physical read io requests' and 'physical write io requests' metrics'
  • (Contrib) exporter/awss3: Implement timeout for S3 exporter (#36264)
  • (Contrib) extension/bearertokenauth: Allow the header name to be customized in the bearerauthtoken extension (#38793)
  • (Contrib) receiver/hostmetrics: Reduced the cost of retrieving number of threads and parent process ID on Windows.
    Disable the featuregate hostmetrics.process.onWindowsUseNewGetProcesses to fallback to the previous[] implementation.
    (#32947, #38589)
  • (Contrib) receiver/hostmetrics: Reduced the CPU cost of collecting the process.handles metric on Windows. (#38886)
    Instead of using WMI to retrieve the number of opened handles by each process
    the scraper now uses the GetProcessHandleCount Win32 API which results in
    reduced CPU usage when the metric process.handles is enabled.
  • (Contrib) pkg/ottl: Enhance the Decode OTTL function to support all flavors of Base64 (#38854)
  • (Contrib) receiver/rabbitmq: Enhance the RabbitMQ receiver to collect and report additional node-level metrics: rabbitmq.node.disk_free, rabbitmq.node.disk_free_limit, rabbitmq.node.disk_free_alarm, rabbitmq.node.disk_free_details.rate, rabbitmq.node.mem_used, rabbitmq.node.mem_limit, rabbitmq.node.mem_alarm, rabbitmq.node.mem_used_details.rate, rabbitmq.node.fd_used, rabbitmq.node.fd_total, rabbitmq.node.fd_used_details.rate, rabbitmq.node.sockets_used, rabbitmq.node.sockets_total, rabbitmq.node.sockets_used_details.rate, rabbitmq.node.proc_used, rabbitmq.node.proc_total, rabbitmq.node.proc_used_details.rate. These provide additional observability into the state and resource usage of RabbitMQ nodes. (#38976)
  • (Contrib) receiver/rabbitmq: Enhance the RabbitMQ receiver to collect and report additional node-level metrics across multiple categories. These include metrics related to memory, file descriptors, sockets, processes, disk, uptime, scheduling, garbage collection (GC), I/O, message store, connections, clustering, configuration, application info, and context switches. This significantly improves visibility into the performance, state, and resource usage of RabbitMQ nodes. (#38997)
  • (Contrib) processor/resourcedetection: Adding the os.version resource attribute to system resourcedetection processor (#38087)
  • (Contrib) receiver/sqlserver: support sqlserverreceiver to record every executing query (#36462)
    We introduced Query Sample collection in this PR. The scraper will record all the currently
    executing queries once (in most case) and report related metrics. With this, user will be
    able to see what queries got executed and can combine with the Top Query to get more insights
    on troubleshooting and fine tuning.
  • (Contrib) receiver/sqlserver: Support query-level log collection (#36462)
    Added top query (most CPU time consumed) collection. The query will gather the queries took most of the time during the last
    query interval and report related metrics. The number of queries can be configured. This will enable user to have better
    understanding on what is going on with the database. This enhancement empowers users to not only monitor but also actively
    manage and optimize their MSSQL database performance based on real usage patterns.

🧰 Bug fixes 🧰

  • (Splunk) Add headers_setter extension to OTLP HTTP exporters in the default configs to facilitate token passthrough (#6083, #6075)
  • (Core) exporters: Fix order of starting between queue and batch. (#12705)
  • (Contrib) receiver/awsecscontainermetrics: Ensure that the storage.read_bytes and storage.write_bytes metrics include i/o counts from all devices (#38301)
  • (Contrib) receiver/sqlserver: the current metric scraper would report error when parsing value, the value was parsed as int but actually it should be a float (#38823)
  • (Contrib) receiver/sqlserver: Ensure all enabled metrics are emitted (#38839)
  • (Contrib) receiver/kafka: make calls to error backoff thread-safe and add logging (#38941)
  • (Contrib) receiver/sqlserver: Remove warning that server.address and server.port resource attributes will be enabled (#38831)
    There is no intention of enabling these resource attributes by default,
    this change is to simply remove the warning.
    The server.address and server.port resource attributes were, and still are,
    disabled by default.
  • (Contrib) pkg/stanza: Add retries when calls to retrieve Windows event via EvtNext fail with error RPC_S_INVALID_BOUND (1734). (#38149)
    Whenever large events were read in by the Windows event log receiver, via the stanza input operator,
    the collector would fail with error RPC_S_INVALID_BOUND (1734). Now the operator tries to workaround
    this issue by reducing the number of events read on each attempt.

Docker Image Manifests:

  • Linux (amd64, arm64, ppc64le) and Windows (2019 amd64, 2022 amd64):
    • quay.io/signalfx/splunk-otel-collector:0.123.0
    • digest: sha256:e576ea17cce918003b8dac7cb625d5e77c3400b4ce974b52a0e64a779429175d
  • Windows (2019 amd64, 2022 amd64):
    • quay.io/signalfx/splunk-otel-collector-windows:0.123.0
    • digest: sha256:fb7a7ab74b02bc9baaec9d68ecfe831d88069b3dffe1c36c41e98f7ace5c8857
Assets 32
Loading