2.7.0 container can't find default paths

[cemsbr]

Bug Report

What did you do?
Upgraded to docker image 2.7.0 from 2.6.1

What did you expect to see?

• Container up
• Preserved collected data

What did you see instead? Under which circumstances?

• Container exited complaining about not finding prometheus.yml (it remained in the same location). Workaround: add the default option --config.file=/etc/prometheus/prometheus.yml to command in docker-compose.yml
• After applying the workaround above, no data was displayed in grafana (maybe the workaround is analogous: specifying the default option in command).

Reverting to v2.6.1 worked.

Environment
Ubuntu 18.04, docker-compose version 1.23.2, Docker version 18.06.1-ce.

• System information:

  Linux 4.15.0-1029-aws x86_64

• Prometheus version:

  2.7.0, from docker hub

• Alertmanager version:

  alertmanager, version 0.16.0

• Prometheus configuration file:

# my global config
global:
  scrape_interval:     15s # Set the scrape interval to every 15 seconds. Default is every 1 minute.
  evaluation_interval: 15s # Evaluate rules every 15 seconds. The default is every 1 minute.
  # scrape_timeout is set to the global default (10s).

# Alertmanager configuration
alerting:
  alertmanagers:
  - static_configs:
    - targets:
      - alertmanager:9093

# Load rules once and periodically evaluate them according to the global 'evaluation_interval'.
rule_files:
- "alert_rules.yml"
  # - "second_rules.yml"

# A scrape configuration containing exactly one endpoint to scrape:
# Here it's Prometheus itself.
scrape_configs:
  # The job name is added as a label `job=<job_name>` to any timeseries scraped from this config.
- job_name: 'prometheus'

  # metrics_path defaults to '/metrics'
  # scheme defaults to 'http'.

  static_configs:
  - targets:
    - 'prometheus:9090'
  relabel_configs:
  - source_labels: [__address__]
    target_label: instance
    regex: ([^\.]+).*:\d+

- job_name: ...

• Logs:
  It was complaining about not finding prometheus.yml

• Docker compose:

  prometheus:
    ports:
      # Access via tunnel (no user/pass via web)
      - 9090:9090
    image: prom/prometheus:v2.6.1
    command:
      - '--storage.tsdb.retention=60d'
    volumes:
      - ./config/prometheus/prometheus.yml:/etc/prometheus/prometheus.yml:ro
      - ./config/prometheus/alert_rules.yml:/etc/prometheus/alert_rules.yml:ro
      - prometheus-storage:/prometheus

[simonpasquier]

As noted in the v2.7.0 changelog, Dockerfile has been reverted to its v2.5.0 version because the changes broke too many users (see ##5043). You probably have to specify both --config.file=/etc/prometheus/prometheus.yml and --storage.tsdb.path=/prometheus.

[cemsbr]

These settings are included in prometheus Dockerfile in the release-2.7.0 branch. Shouldn't it just work?

[cemsbr]

Probably my commands config overrode the CMD in Dockerfile, as stated in https://docs.docker.com/compose/compose-file/#command.