Add process metadata

edit

The add_process_metadata processor enriches events with information from running processes, identified by their process ID (PID).

processors:
  - add_process_metadata:
      match_pids:
        - process.pid

The fields added to the event look as follows:

{
  "container": {
    "id": "b5285682fba7449c86452b89a800609440ecc88a7ba5f2d38bedfb85409b30b1"
  },
  "process": {
    "args": [
      "/usr/lib/systemd/systemd",
      "--switched-root",
      "--system",
      "--deserialize",
      "22"
    ],
    "executable": "/usr/lib/systemd/systemd",
    "name": "systemd",
    "owner": {
      "id": "0",
      "name": "root"
    },
    "parent": {
      "pid": 0
    },
    "pid": 1,
    "start_time": "2018-08-22T08:44:50.684Z",
    "title": "/usr/lib/systemd/systemd --switched-root --system --deserialize 22"
  }
}

Optionally, the process environment can be included, too:

  ...
  "env": {
    "HOME":       "/",
    "TERM":       "linux",
    "BOOT_IMAGE": "/boot/vmlinuz-4.11.8-300.fc26.x86_64",
    "LANG":       "en_US.UTF-8",
  }
  ...

It has the following settings:

match_pids
List of fields to lookup for a PID. The processor will search the list sequentially until the field is found in the current event, and the PID lookup will be applied to the value of this field.
target
(Optional) Destination prefix where the process object will be created. The default is the event’s root.
include_fields
(Optional) List of fields to add. By default, the processor will add all the available fields except process.env.
ignore_missing
(Optional) When set to false, events that don’t contain any of the fields in match_pids will be discarded and an error will be generated. By default, this condition is ignored.
overwrite_keys
(Optional) By default, if a target field already exists, it will not be overwritten, and an error will be logged. If overwrite_keys is set to true, this condition will be ignored.
restricted_fields
(Optional) By default, the process.env field is not output, to avoid leaking sensitive data. If restricted_fields is true, the field will be present in the output.
host_path
(Optional) By default, the host_path field is set to the root directory of the host /. This is the path where /proc is mounted. For different runtime configurations of Kubernetes or Docker, the host_path can be set to overwrite the default.
cgroup_prefixes
(Optional) List of prefixes that will be matched against cgroup paths. When a cgroup path begins with a prefix in the list, then the last element of the path is returned as the container ID. Only one of cgroup_prefixes and cgroup_rexex should be configured. If neither are configured then a default cgroup_regex value is used that matches cgroup paths containing 64-character container IDs (like those from Docker, Kubernetes, and Podman).
cgroup_regex
(Optional) A regular expression that will be matched against cgroup paths. It must contain one capturing group. When a cgroup path matches the regular expression then the value of the capturing group is returned as the container ID. Only one of cgroup_prefixes and cgroup_rexex should be configured. If neither are configured then a default cgroup_regex value is used that matches cgroup paths containing 64-character container IDs (like those from Docker, Kubernetes, and Podman).
cgroup_cache_expire_time
(Optional) By default, the cgroup_cache_expire_time is set to 30 seconds. This is the length of time before cgroup cache elements expire in seconds. It can be set to 0 to disable the cgroup cache. In some container runtimes technology like runc, the container’s process is also process in the host kernel, and will be affected by PID rollover/reuse. The expire time needs to set smaller than the PIDs wrap around time to avoid wrong container id.