Monitoring and logging

StatisticsStatistics

Load balancer statistics are automatically logged in the Yandex Monitoring metrics. The following dashboards and measures are available:

• HTTP statistics:

  • RPS: Number of load balancer requests per second.
  • 4XX, 5XX: Number of load balancer responses containing HTTP codes 4XX and 5XX and the corresponding gRPC codes per second.
  • Request size: Total volume of load balancer requests per second.
  • Response size: Total volume of load balancer responses per second.
  • Latency: Response delay (the time between the balancer receiving the first byte of a request to sending the last byte of the response), 50th to 99th percentiles.

• Capacity statistics:

  • Active connections: Number of active connections.
  • Connections per second: Number of connections per second.
  • Requests per second: Number of requests per second.
  • Bytes per second: Amount of data handled per second.

For a full list of metrics delivered to Yandex Monitoring, see the reference.

Application Load Balancer has aggregate load balancer statistics available. In Monitoring , you can view statistics itemized by the resources associated with the load balancer (HTTP routers, virtual hosts, routes, and the like) as well as create alerts.

For instructions on viewing statistics, see Viewing L7 load balancer statistics.

LoggingLogging

You can configure the delivery of load balancer logs to a Yandex Cloud Logging log group.

For more information on how to view logs, see Viewing L7 load balancer logs.

The X-Forwarded-For (XFF) header value is logged as per RFC 7239. You can find the full list of logged parameters in the log reference.

You can also send load balancer logs to a PostgreSQL DB.

Log discard rulesLog discard rules

Writing and storing logs in Cloud Logging is charged based on the service pricing rules. To log less data, add log discard rules.

Possible rules:

See alsoSee also

• Monitoring charts
• Setting up alerts