Stackdriver Config
阅读大约需要 3 分钟
CustomConfig
Custom instance configuration overrides. Provides a way to customize metrics/logs.
Field | Type | Description | Required |
---|---|---|---|
dimensions | map<string, string> | (Optional) Collection of tag names and tag expressions to include in the instance. Conflicts are resolved by the tag name by overriding previously supplied values. | No |
PluginConfig
next id: 15
Field | Type | Description | Required |
---|---|---|---|
max_log_batch_size_in_bytes | int32 | Optional. Allows configuration of the size of the LogWrite request. The size is in bytes, so that it allows for better performance. Default is 4MB. The size of one log entry within LogWrite request is approx 1Kb. | No |
log_report_duration | Duration | Optional. Allows configuration of the time between calls out to the stackdriver logging service to report buffered LogWrite request. Customers can choose to report more aggressively by keeping shorter report interval if needed. Default is 10s. | No |
enable_audit_log | bool | Optional. Controls whether to export audit log. | No |
destination_service_name | string | Optional. FQDN of destination service that the request routed to, e.g. productpage.default.svc.cluster.local. If not provided, request host header will be used instead | No |
enable_mesh_edges_reporting | bool | Optional. Controls whether or not to export mesh edges to a mesh edges service. This is disabled by default. | No |
mesh_edges_reporting_duration | Duration | Optional. Allows configuration of the time between calls out to the mesh
edges service to report NEW edges. The minimum configurable duration is
| No |
max_peer_cache_size | int32 | maximum size of the peer metadata cache. A long lived proxy that connects with many transient peers can build up a large cache. To turn off the cache, set this field to a negative value. | No |
disable_host_header_fallback | bool | Optional: Disable using host header as a fallback if destination service is not available from the controlplane. Disable the fallback if the host header originates outsides the mesh, like at ingress. | No |
max_edges_batch_size | int32 | Optional. Allows configuration of the number of traffic assertions to batch into a single request. Default is 100. Max is 1000. | No |
disable_http_size_metrics | bool | Optional. Allows disabling of reporting of the request and response size metrics for HTTP traffic. Defaults to false (request and response size metrics are enabled). | No |
enable_log_compression | BoolValue | Optional. Allows enabling log compression for stackdriver access logs. | No |
access_logging | AccessLogging | Optional. Controls what type of logs to export.. | No |
custom_log_config | CustomConfig | (Optional) Collection of tag names and tag expressions to include in the logs. Conflicts are resolved by the tag name by overriding previously supplied values. Does not apply to audit logs. See https://istio.io/latest/docs/tasks/observability/metrics/customize-metrics/#use-expressions-for-values for more details about the expression language. | No |
metric_expiry_duration | Duration | Optional. Controls the metric expiry duration. If a metric time series is not updated for the given duration, it will be purged from time series cache as well as metric reporting. If this is not set or set to 0, time series will never be expired. This option is useful to avoid unbounded metric label explodes proxy memory. | No |
disable_server_access_logging | bool | Optional. Controls whether to export server access log. This is deprecated in favor of AccessLogging enum. | No |
PluginConfig.AccessLogging
Types of Access logs to export. Does not affect audit logging.
Name | Description |
---|---|
NONE | No Logs. |
FULL | All logs including both success and error logs. |
ERRORS_ONLY | All error logs. This is currently only available for outbound/client side
logs. A request is classified as error when |