Log Collection

When a user wants to download logs from the Edge Device UI (how to download logs), Device Builders must ensure that logs are collected and placed in a specified folder.

Responsibilities of IEDK GetLogFile rpc Method

1. Collect journalctl outputs
2. If additional log files exist that do not use journalctl, these log files can be added (optional)
3. Compress the log files and place the archived file to the requested directory
4. Return the full path of the log file

Log Collection Scenario

The Edge Device UI should call the GetLogFile method with an existing directory name as a request parameter, the GetLogFile rpc method must handle the log collection, compression, naming etc.

You can get whole system logs from journalctl with the following command:

journalctl > logs

Or just add specific logs:

journalctl -u networking > logs

If you want, you can add custom logs as well.

Let's assume that you have a log file named myCustomLogFile.

cat myCustomLogFile >> logs

You need to compress the collected log file and put it under the requested directory.

NOTICE

If the requested directory does not exist, the service need to return an error because the creation of the directory is not the responsiblity of GetLogFile rpc method.

  tar -czvf /requested_path/devicelogs_devicename_timestamp.tar.gz logs

GetLogFile rpc method must return the full log file path as response. For the case above, /requested_path/devicelogs_devicename_timestamp.tar.gz should be returned.

Log File Naming

There are no restrictions on the naming of the collected logs. It is recommended to use a unique identifier in the log name that describes the device individually, for example: devicelogs_devicename_timestamp.tar.gz

Custom Log Collection and Forwarding to Third-party IT Infrastructure

Device Builders can set up the log collector by modifying the host.log.driver capability in the capabilities.json. This configuration allows users to efficiently route collected logs to their specified IT infrastructure. Available collector options include:

• syslog: Create a /var/run/syslog.sock socket file in edge iot core to accept log data.
• systemd: Reads log data from the /host-system/var/log/journal folder, and it serves as the default log collector.
• tail: Reads log data from the /data/agent/tail-log file.
• none: No logs will be collected and forwarded.

Capability Example

Below is a simplified capabilities.json configuration example, setting only the host.log.driver capability:

{
  "capabilityId": "host.log.driver",
  "name": "Select log driver for device",
  "description": "Only one driver can be selected, systemd is the default selection.",
  "scope": "protected",
  "property": "optional",
  "details": {
    "exLogDriver": "systemd"
  }
}

Deploying via docker compose

It is important to mount files or directories to different log collectors. Below is a simplified docker-compose.yml example that shows setups for different types of log collectors:

version: "2.4"
services:
  industrial-edge:
    restart: always
    environment:
      - HostIps=X.X.X.X
    volumes:
      - "/etc/capabilities.json:/data/app_engine/Configuration/capabilities.json"
      - "/host-system/xxx/journal:/host-system/var/log/journal" # systemd collector
      - "/host-system/xxx:/var/run" # syslog collector
      - "/host-system/xxx/tail-log:/data/agent/tail-log" # tail collector
    ports:
      - "443:443"
    container_name: edge-iot-core
    image: edge_iot_core:1.16.0

Forwarding to Third-party IT Infrastructure

Logs can be forwarded to third-party IT infrastructures via Logging & Monitoring. For example, to forward system logs to AWS S3:

1. Log in to the IED UI, navigate to Settings > Logging & Monitoring, and Creating Configuration to forward the system log to AWS S3.

2.After configuration, access your AWS S3 service to verify that the system log was successfully forwarded to the specified bucket.