Troubleshooting OpenTelemetry with New Relic

Troubleshooting OpenTelemetry with New Relic may just be a matter of making sure you are following best practices, but sometimes you may need to take additional steps to diagnose your issues. Here are some examples of specific problems you might encounter, along with steps and tools to resolve them.

OpenTelemetry data sent via OTLP is not queryable

Problem

You sent OpenTelemetry metrics, logs, or traces using OTLP and are unable to view the data. Before digging deeper, make sure you've checked the following:

• The OTLP endpoint configured matches one of our documented endpoints, is properly formatted, and includes the default port for OTLP/gRPC, 4317, or the default port for OTLP/HTTP, 4318. Port 443 is also supported for either transport. Please note the specific endpoint for FedRAMP compliance, if applicable.
• Confirm that you are using the correct endpoint based on your region. For example, if you are based in Europe and configure your exporter to send data to our US endpoint, data will fail to be exported as the endpoints are region-specific.
• The outbound traffic is not restricted by a firewall. Our Networks document explains domains and network blocks that you may need to explicitly allow.
• The client is configured to use TLS 1.2 or higher and the request includes the api-key header with a valid New Relic account (ingest) license key.
• Requests include valid protobuf payloads and use gRPC or HTTP transport, preferably with gzip compression enabled. Request payloads should be no larger than 1MB (10^6 bytes). JSON-encoded protobuf payloads are also supported.
• Client output and logs do not indicate 4xx or 5xx response codes are being returned.

Solution

There are number of tools you can use to validate the successful delivery of telemetry data to our platform. A good first step is to check the data management hub to facet data ingest and determine how much data is arriving from various sources. You can also use the data explorer or query builder to look for data faceted by instrumentation.provider or newrelic.source attributes:

FROM Log, Metric, Span SELECT datapointcount() WHERE instrumentation.provider = 'opentelemetry' FACET instrumentation.provider, newrelic.source

This query should tell you whether data is arriving via OTLP. If the data you expect is not present, try this alternate query:

FROM Log, Metric, Span SELECT count(*) where newrelic.source LIKE 'api.%.otlp'

You can also check for integration errors by querying NrIntegrationError events. This can help you determine whether you have configuration or format issues or if you've run into our platform limits.

Various parts of the New Relic UI rely on the presence of specific attributes to function properly. You can use the NRQL console feature in many places to check the WHERE or FACET clauses of the query for required attributes. You can also edit those clauses and re-run the query to determine whether there is data present with those attributes missing. Examples of required attributes include service.name and service.instance.id. For a more complete list of examples, see resources.

OpenTelemetry entities or relationships are missing

Problem

You sent OpenTelemetry data from a service or infrastructure component and either the entity or its relationships are missing or incorrect.

Solution

OpenTelemetry entities will be synthesized based on the public rules described for the EXT-SERVICE entity type. The standard rule to match relies on the presence of the service.name dimension which follows the OpenTelemetry semantic conventions.

To set the service.name with the OpenTelemetry Java SDK, include it in your resource:

var resource = Resource.getDefault()
    .merge(Resource.builder().put(SERVICE_NAME, serviceName).build());

Depending on the SDK, you may also set the service.name by declaring it in the OTEL_RESOURCE_ATTRIBUTES or OTEL_SERVICE_NAME environment variables.

For log management, you can use a structured log template to inject the service.name. Here are some log examples:

• Setting the service name
• Logs in context with Log4j2

Troubleshooting the OpenTelemetry collector

The best place for collector troubleshooting tips and monitoring practices is the up-to-date guidelines in the OpenTelemetry community. See the links below for community troubleshooting documents.

Collector logs

Set the log level in the config service::telemetry::logs. The default level is INFO. Supported levels are: DEBUG, INFO, WARN, ERROR, DPANIC, PANIC, FATAL.

For troubleshooting tips, see logs troubleshooting (GitHub).

Collector metrics

The following NRQL query shows all the available metrics from the collector itself in New Relic:

FROM Metric SELECT uniques(metricName) WHERE metricName like ‘otelcol_%’ LIMIT MAX

For troubleshooting tips, see:

• Metric troubleshooting (GitHub)
• Monitoring collector metrics (GitHub)

Collector traces

For troubleshooting tips, see zPages (Github).