-
Notifications
You must be signed in to change notification settings - Fork 494
Open
Labels
Description
Component(s)
otelcol.connector.servicegraph
What's wrong?
When sending exponential histograms to Grafana cloud the client metric traces_service_graph_request_client_seconds for the service graph fails to send because of no timestamp
The traces_service_graph_request_server_seconds has no issues and is is successfully sent.
Steps to reproduce
Use the following in your alloy config
otelcol.connector.servicegraph "default" {
dimensions = [
"service.namespace",
"service.version",
"deployment.environment",
"k8s.cluster.name",
"k8s.namespace.name",
"team",
]
exponential_histogram_max_size = 20
store {
ttl = "2s"
}
output {
metrics = [otelcol.processor.transform.trace_metrics.input]
}
}
System information
K8s 1.33.*
Software version
Grafana Alloy 1.12.0
Configuration
Logs
ts=2025-12-04T18:10:02.156939816Z level=error msg="Exporting failed. Dropping data." component_path=/ component_id=otelcol.exporter.otlphttp.lgtm error="not retryable error: Permanent error: rpc error: code = InvalidArgument desc = error exporting items, request to https://otlp-gateway-****.grafana.net/otlp/v1/metrics responded with HTTP Status Code 400, Message=received a sample whose timestamp is too far in the past, timestamp: 0 series: 'traces_service_graph_request_client_seconds' (err-mimir-too-far-in-past). To adjust the related per-tenant limit, configure -validation.past-grace-period, or contact your service administrator., Details=[]" dropped_items=407
Tip
React with 👍 if this issue is important to you.