Studio Conductor System Processes and Logging Guide

The purpose of this document is to provide detailed information for:

• Identifying Conductor Manager’s processes
• Defining how these processes should be tracked for monitoring and alerting
• Defining locations of Conductor Manager log files

Studio Conductor System Processes

In a Conductor Manager environment, the following system processes exist:

Studio Conductor Systemd Init Services

Conductor Service Configuration Defaults

All Studio Conductor-specific service configurations can be found in /etc/sysconfig. This area is where default configuration data can be found as well as logging locations for service-specific troubleshooting. These are very useful when trying to understand how a service was instantiated and what logging configuration is being used.

This directory can also be used to derived each core service’s Systemd init name. For instance, enumerating /etc/sysconfig will show a file called cloudify-stage. This is the name of the service, and thus to query the service status can be done using the command service cloudify-stage status.

Discovering Studio Conductor Services

Service Statuses

The sections above describe how to identify a Studio Conductor service by looking directly at the output
of something like ps or by folder snooping. This is not always practical or desired and there are
other, more developer-friendly, ways of enumerating which Studio Conductor services are present and
how to harvest information about them.

The best starting point is to utilize the REST API of the manager to get service information.
Simply craft a GET request for the status endpoint: GET /status HTTP/1.1
If cURL and Python are available, it’s very easy to make the request as well as visualize the
returned data.

Code Block 1 REST

curl -X GET http://<manager-ip>/status | python -m json.tool

An example, partial, return is as follows:

Code Block 2 JSON

{
"services": [{
"display_name": "RabbitMQ",
"instances": [{
"ActiveState": "active",
"Description": "RabbitMQ Service",
"Id": "cloudify-rabbitmq.service",
"LoadState": "loaded",
"MainPID": 12322,
"SubState": "running",
"state": "running"
}
]
}

With this information, in standard JSON format, it is easy to match a core Studio Conductor service with a
system-level process ID (MainPID) to begin further troubleshooting.

Cluster status

Studio Conductor provides system health information for both single box deployments and clustered deployments. Read more about it:

• Cluster Status widget
• Cluster status
• Manager status

Checking Manager Components

RabbitMQ

System Service

To check if the RabbitMQ broker is running (and to see many other details such as which applications are running, memory allocation, and other performance metrics), simply run the following command:

Code Block 7 bash

sudo rabbitmqctl -n cloudify-manager@localhost status

An error message will be presented if the service has an issue such as a failed broker.

Management Operations

To get started working with the RabbitMQ management interface, the management interface must be enabled via a plugin. Execute the following to enable the management plugin:

sudo rabbitmq-plugins -n cloudify-manager@localhost enable rabbitmq_management

Once this is complete, there will be a management web interface located at http://:15672/

In order to utilize the web interface, you will need to have the RabbitMQ username and password for authentication. This can be found in the /etc/cloudify/config.yaml file used for instantiation a Conductor Manager.

By default, the user created from the manager instantiation process does not have sufficient permissions to be used with the web interface. Use the following command to promote the default user with the “monitoring” permission (or you can alternatively assign the “administrator” tag).

sudo rabbitmqctl set_user_tags <username> monitoring

You can now use the RabbitMQ username and password to log in via the web interface to do actions such as view queues, get messages, monitor performance, and monitor connections.

PostgreSQL

System Service

To verify if postgres is working correctly a simple select can be executed:

Code Block 10 bash

sudo -u postgres psql --port 15432 -c "select 1"

Syncthing

Checking if syncthing is working correctly will need a curl command to the REST API.

Code Block 12 REST

curl -H "X-Api-Key: <key>" 127.0.0.1:8384/rest/system/status

The key can be gathered from: //configuration/gui/apikey in /opt/syncthing/.config/syncthing/config.xml

Logging

Overview

Log locations vary from service to service, but the majority of logs can be found in /var/log and /var/log/cloudify.

Within these folders are folders for each service with distinguishable names such as “rabbitmq” and “postgres”. If logs for a service aren’t found here, the next place to look would be in the service configuration defaults file for any indication of a log file path (see the section “Studio Conductor Service Configuration Defaults”).

Studio Conductor Agent Worker Logs

Studio Conductor agent worker logs can be found on deployed instances / virtual machines with an installed Studio Conductor agent. Typically, the logs are stored in the Studio Conductor agent user’s home directory in a folder named after the node instance ID for the instance / VM.

• The Celery service SysV Init file is /etc/init.d/celeryd-.
• The Celery service config file is /etc/default/celeryd-.
• Studio Conductor agent worker log. ~//work/.log
• This is the agent counterpart to the Studio Conductor Management Worker logs. ~//work/-.log
• Worker-specific log.
  • Each worker gets its own numbered log file. ~//work/%I.log

Studio Conductor Management Worker Logs

Studio Conductor Management Worker Logs

• /var/log/cloudify/mgmtworker/cloudify.management_worker.log

  • Studio Conductor management worker log.
  • Useful for troubleshooting management worker issues such as Studio Conductor agent deployment, blueprint deployment creation, and heartbeat errors.
  • Contains information about deployment executions from the perspective of the management worker.
  • Shows worker tracebacks.
  • Task execution logs are followed by noting the task dispatch ID (a UUID). Task IDs can also be found in execution logs and used to search this worker log for further details. Specific task logs will have prefixes of “Received task”, “Task accepted”, and “Task [succeeded | failed]“. Here’s an example:

  Code Block 13 LOG

  Received task: cloudify.dispatch.dispatch[b164cf2c-d601-4484-bbce-927e1106de27]
  Task accepted: cloudify.dispatch.dispatch[b164cf2c-d601-4484-bbce-927e1106de27] pid:5683
  Task cloudify.dispatch.dispatch[b164cf2c-d601-4484-bbce-927e1106de27] succeeded in 1.015225859s

• /var/log/cloudify/mgmtworker/logs/.log

  • Conductor deployment worker log.
  • Useful for troubleshooting deployment executions of all types. Low-level logging of worker tasks and is generally used as an additional source of information if the execution logs themselves aren’t sufficient.
  • Shows worker tracebacks.

Studio Conductor REST API Service Logs

• /var/log/cloudify/rest/cloudify-rest-service.log

  • Serves as a central log file for all incoming and outgoing REST API requests and responses. Log entries are in a well-defined, human-readable format.
  • Provides a host of useful information such as request details (HTTP method, headers, query string details, JSON data, endpoint path, etc…) and response details (HTTP status, headers).
  • Can be monitored, on-demand, for bad HTTP response codes, blueprint file names, endpoint security checks, etc

• /var/log/cloudify/rest/gunicorn-access.log

  • Verbose access logs directly from the HTTP server itself.
  • Well-structured, dense logging format.
  • Useful for monitoring REST API interaction, user fingerprinting, and this log file includes maintenance endpoint calls and other “internal” endpoints that Studio Conductor uses.

• /var/log/cloudify/rest/gunicorn.log

  • Gunicorn HTTP server system service log.
  • Useful for troubleshooting SysV init service failures as well as enumerating the
  • HTTP server worker process IDs and HTTP server listening endpoint.

PostgreSQL Logs

• /var/log/cloudify/postgresql

  • PostgreSQL system service log.
  • Useful for gathering information about the PostgreSQL service such as version, process ID, build, and cluster information.
  • Useful for monitoring cluster state and indexing tasks.
  • Useful for PostreSQL service troubleshooting.

RabbitMQ Logs

• /var/log/cloudify/rabbitmq/.log

  • RabbitMQ system service log.
  • Useful for gathering information about the RabbitMQ service such as node name, config file locations, database directory, and running reporting info.
  • Useful for RabbitMQ service troubleshooting.

Cluster Logs

• /var/log/cloudify/cloudify-cluster.log

  • Cluster services log.
  • All cluster services log to this file and journald.
  • Useful for gathering information about Cluster operations.