The pmm-managed.log is continuously printing error logs regarding evaluation_interval option

When I set the evaluation_interval option in the /srv/prometheus/prometheus.base.yml file, the pmm-managed.log is continuously printing error logs regarding evaluation_interval option.

Here is a content of the file.

# cat /srv/prometheus/prometheus.base.yml
global:
  evaluation_interval: 10s
scrape_configs:
- job_name: aaa
  honor_timestamps: false
  scrape_interval: 10s
  scrape_timeout: 9s
  scheme: https
  metrics_path: /metrics
  static_configs:
  - targets:
    - aaa:8443
  tls_config:
    insecure_skip_verify: true

Here is the logs.

ERRO[2022-08-23T17:54:44.888-07:00] 2022-08-24T00:54:44.880Z    info    /home/builder/rpm/BUILD/Victor                                                                                iaMetrics-pmm-6401-v1.77.1/lib/logger/flag.go:12        build version: victoria-metrics-20220721-12201                                                                                9-pmm-6401-v1.77.1
2022-08-24T00:54:44.880Z        info    /home/builder/rpm/BUILD/VictoriaMetrics-pmm-6401-v1.77.1/lib/l                                                                                ogger/flag.go:13        command line flags
2022-08-24T00:54:44.880Z        info    /home/builder/rpm/BUILD/VictoriaMetrics-pmm-6401-v1.77.1/lib/l                                                                                ogger/flag.go:20        flag "dryRun"="true"
2022-08-24T00:54:44.880Z        info    /home/builder/rpm/BUILD/VictoriaMetrics-pmm-6401-v1.77.1/lib/l                                                                                ogger/flag.go:20        flag "enableTCP6"="true"
2022-08-24T00:54:44.880Z        info    /home/builder/rpm/BUILD/VictoriaMetrics-pmm-6401-v1.77.1/lib/l                                                                                ogger/flag.go:20        flag "promscrape.config"="/tmp/pmm-managed-config-victoriametrics-2205029429"
2022-08-24T00:54:44.887Z        fatal   /home/builder/rpm/BUILD/VictoriaMetrics-pmm-6401-v1.77.1/app/v                                                                                ictoria-metrics/main.go:54      error when checking -promscrape.config: cannot parse Prometheus config                                                                                 from "/tmp/pmm-managed-config-victoriametrics-2205029429": cannot unmarshal data: yaml: unmarshal err                                                                                ors:
  line 6: field evaluation_interval not found in type promscrape.GlobalConfig; pass -promscrape.config                                                                                .strictParse=false command-line flag for ignoring unknown fields in yaml config  component=victoriamet                                                                                rics
ERRO[2022-08-23T17:54:44.888-07:00] Failed to update configuration, will retry: rpc error: code = Abor                                                                                ted desc = yaml: unmarshal errors:
  line 6: field evaluation_interval not found in type promscrape.GlobalConfig; pass -promscrape.config                                                                                .strictParse=false command-line flag for ignoring unknown fields in yaml config
.  component=victoriametrics

I’m using the PMM Server 2.29.1 version.

# pmm-admin --version
ProjectName: pmm-admin
Version: 2.29.1
PMMVersion: 2.29.1
Timestamp: 2022-07-25 07:02:19 (UTC)
FullCommit: 2c740d96c9399a5140f5a16b2063024b895753c4

How can I stop it?

Also, in that situation, the content of the /srv/prometheus/prometheus.base.yml file did not get merged into the /etc/victoriametrics-promscrape.yml file too.

Your error log says evaluation_interval is an invalid option. Can you try removing that and try again?

Thank you for your reply. But it seems that a user can set the option and it is referenced by below code block for vmalert’s evaluationInterval parameter.

I just want to set the evaluationInterval for alerts. When I run the PMM 2.23.0 version with such option, it worked fine and the vmalert configured correctly with evaluationInterval parameter.

Hi. Are there any updates? Thank you for your supports in advance. :slight_smile: