このページは日本語には対応しておりません。随時翻訳に取り組んでいます。翻訳に関してご質問やご意見ございましたら、お気軽にご連絡ください。
The Datadog Agent logs pipeline is enabled by default in the Datadog Exporter in v0.108.0. This may cause a breaking change if
logs::dump_payloads is in use while upgrading, since this option is invalid when the Datadog Agent logs pipeline is enabled. To avoid this issue, remove the
logs::dump_payloads config option or temporarily disable the
exporter.datadogexporter.UseLogsAgentExporter feature gate.
Overview
To collect logs from files, configure the filelog receiver in your Datadog Exporter.
For more information, see the OpenTelemetry project documentation for the filelog receiver.
Setup
For a collector deployed on the same host as the log files to be collected, specify the paths of the log files to collect in your Collector configuration:
receivers:
filelog:
include_file_path: true
poll_interval: 500ms
include:
- /var/log/*/app.log
operators:
- type: json_parser
# Layout must match log timestamp format. If this section is removed, timestamp will correspond to the time of log intake by Datadog.
- type: time_parser
parse_from: attributes.time
layout: '%Y-%m-%dT%H:%M:%S%z'
Add the following lines to values.yaml:
presets:
logsCollection:
enabled: true
includeCollectorLogs: true
The filelog receiver needs access to the file paths. The preset mounts the necessary volumes to the collector container for /var/log/pods and collects all logs from /var/log/pods/*/*/*.log. See Important components for Kubernetes for a full list of settings set by the preset.
Collector configuration sets up a list of operators to parse the logs based on different formats:
filelog:
include:
- /var/log/pods/*/*/*.log
exclude:
- /var/log/pods/abc/*.log
operators:
- type: json_parser
- type: trace_parser
trace_id:
parse_from: attributes.trace_id
span_id:
parse_from: attributes.span_id
trace_flags:
parse_from: attributes.trace_flags
- type: time_parser
parse_from: attributes.time
layout: '%Y-%m-%dT%H:%M:%S%z'
In order to add custom Datadog tags to logs, set the ddtags attribute on the logs. For example, this can be done with the transform processor:
processors:
transform:
log_statements:
- context: log
statements:
- set(attributes["ddtags"], "first_custom:tag, second_custom:tag")
Data collected
Logs from the configured files.
Full example configuration
For a full working example configuration with the Datadog exporter, see logs.yaml.
Example logging output
ResourceLog #0
Resource SchemaURL: https://opentelemetry.io/schemas/1.6.1
Resource attributes:
-> k8s.container.name: Str(loadgenerator)
-> k8s.namespace.name: Str(otel-staging)
-> k8s.pod.name: Str(opentelemetry-demo-loadgenerator-d8c4d699d-ztt98)
-> k8s.container.restart_count: Str(1)
-> k8s.pod.uid: Str(92bf09ed-0db9-4f69-a9d6-1dadf12e01aa)
-> k8s.pod.ip: Str(192.168.55.78)
-> cloud.provider: Str(aws)
-> cloud.platform: Str(aws_ec2)
-> cloud.region: Str(us-east-1)
-> cloud.account.id: Str(XXXXXXXXX)
-> cloud.availability_zone: Str(us-east-1c)
-> host.id: Str(i-0368add8e328c28f7)
-> host.image.id: Str(ami-08a2e6a8e82737230)
-> host.type: Str(m5.large)
-> host.name: Str(ip-192-168-53-115.ec2.internal)
-> os.type: Str(linux)
-> k8s.daemonset.uid: Str(6d6fef61-d4c7-4226-9b7b-7d6b893cb31d)
-> k8s.daemonset.name: Str(opentelemetry-collector-agent)
-> k8s.node.name: Str(ip-192-168-53-115.ec2.internal)
-> kube_app_name: Str(opentelemetry-collector)
-> kube_app_instance: Str(opentelemetry-collector)
-> k8s.pod.start_time: Str(2023-11-20T12:53:23Z)
ScopeLogs #0
ScopeLogs SchemaURL:
InstrumentationScope
LogRecord #0
ObservedTimestamp: 2023-11-20 13:02:04.332021519 +0000 UTC
Timestamp: 2023-11-20 13:01:46.095736502 +0000 UTC
SeverityText:
SeverityNumber: Unspecified(0)
Body: Str( return wrapped_send(self, request, **kwargs))
Attributes:
-> log.file.path: Str(/var/log/pods/otel-staging_opentelemetry-demo-loadgenerator-d8c4d699d-ztt98_92bf09ed-0db9-4f69-a9d6-1dadf12e01aa/loadgenerator/1.log)
-> time: Str(2023-11-20T13:01:46.095736502Z)
-> logtag: Str(F)
-> log.iostream: Str(stderr)
Trace ID:
Span ID:
Flags: 0
