Skip to content

[beats receivers] host parsing failed for prometheus-collector: error parsing URL: parse "http://localhost:EDOT_COLLECTOR_METRICS_PORT": invalid port ":EDOT_COLLECTOR_METRICS_PORT" after host #11169

@cmacknz

Description

@cmacknz
  • Version: 9.2.1
  • Operating System: Windows Server 2019 Standard

This was supposed to be fixed in #10819 but we have a bug report that this is still happening in 9.2.1 now that the otel runtime is the default for self-monitoring in 9.2.1

Relevant section of state.yaml in the provided diagnostics:

fleet_message: Connected
fleet_state: 2
log_level: info
message: 1 or more components/units in a failed state
state: 3
components:
    # ...
    - id: prometheus/metrics-monitoring
      state:
        component:
            apmconfig: null
            limits:
                gomaxprocs: 0
                source:
                    fields:
                        go_max_procs:
                            kind:
                                numbervalue: 0
        component_idx: 2
        features_idx: 2
        message: 'Healthy: communicating with pid ''10484'''
        pid: 0
        state: 2
        units:
            input-prometheus/metrics-monitoring-metrics-monitoring-collector:
                message: 'host parsing failed for prometheus-collector: error parsing URL: parse "http://localhost:EDOT_COLLECTOR_METRICS_PORT": invalid port ":EDOT_COLLECTOR_METRICS_PORT" after host'
                payload:
                    streams:
                        metrics-monitoring-collector:
                            error: ""
                            status: STARTING
                state: 4
            output-prometheus/metrics-monitoring:
                message: Healthy
                state: 2
        version_info:
            build_hash: 6485e11edf8854d7792dfc2999bf19db37315ea4
            meta:
                build_time: 2025-11-06 08:49:58 +0000 UTC
                commit: 6485e11edf8854d7792dfc2999bf19db37315ea4
            name: beat-v2-client

Interestingly the prometheus/metrics-monitoring input here is running as a sub-process. I think this is likely because we fell back to using the process runtime because the user is using the Kafka output:

agent:
    download:
        sourceURI: https://artifactory.skoda.vwgroup.com/artifactory/chck-elastic-external-generic-local/agents/
    features: null
    monitoring:
        enabled: true
        logs: true
        metrics: true
        traces: false
        use_output: 60ada9f0-7907-11ee-bc53-e16d9f2c55ad
# ...
outputs:
    60ada9f0-7907-11ee-bc53-e16d9f2c55ad:
        # ...
        required_acks: 1
        sasl:
            mechanism: SCRAM-SHA-512
        ssl:
            verification_mode: full
        timeout: 30
        type: kafka
        version: 2.6.0

Metadata

Metadata

Assignees

Labels

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions