• Русский

Configuring metrics

CodeScoring stores metrics in a format supported by the Prometheus monitoring tool. Currently there are metrics available from CodeScoring installation and plugins in proxy repositories.

Collect metrics from installation

Metrics from installation are available via CodeScoring API at {installation-url}/api/metrics. In order to configure metrics tracking in Prometheus, the following steps are required:

1. open the prometheus.yml configuration file and add parameters to monitor metrics. The following is an example:

   global:
     scrape_interval: 15s
   
     scrape_configs:
     - job_name: 'demo-codescoring'
       metrics_path: '/api/metrics'
       static_configs:
         - targets: ['{installation-url}'] # Installation host address
     - job_name: 'osa'
       metrics_path: '/api/osa/metrics'
       static_configs:
         - targets: ['{installation-url}'] # Installation host address
   

2. Restart Prometheus for the changes to take effect.

3. Open the Prometheus interface and go to the Graph page. In the query field, enter one of the metrics:

Queue Metrics

• codescoring_tasks_queue_size_total - total number of tasks pending;
• codescoring_tasks_schedule_tasks_schedule_queue_size_total - total number of scheduled tasks pending;
• codescoring_tasks_running_tasks_total - total number of tasks currently running.

These metrics can be filtered by the queue label with the following possible values: ipcs, high-priority, osa-container-image-scan , osa-package-scan, policy and tqi.

Analysis Metrics

• codescoring_running_analyses_total is the total number of running analyses. This metric can be filtered by the analysis_type label with the following possible values: sca, authors and clones.

Query metrics from OSA plugins

• codescoring_registration_packages_queue_size - the queue for registering packets from OSA;
• codescoring_registration_registration_container_images_queue_size - queue for registering container images from OSA.

Example of metrics visualization:

Collecting OSA metrics

OSA metrics contain the number and time of requests from plugins, as well as the number of scanned and blocked components in repository managers. The metrics are available via CodeScoring API at {installation-url}/api/osa/metrics.

The following OSA metrics are available:

Connection Metrics.

• codescoring_osa_api_db_connection_pool - database connection pool. This metric can be filtered by the measure label with the following possible values: pool_min, pool_max, pool_size, pool_available and requests_waiting.

• codescoring_osa_api_redis_connection_pool - the pool of connections to Redis. This metric can be filtered by the measure label with the following possible values: available connections, in_use_connections, total_connections.

Component Query Metrics

• codescoring_osa_api_http_request_duration_seconds_sum - total request time;
• codescoring_osa_api_http_request_duration_seconds_bucket - request time by time series;
• codescoring_osa_api_http_requests_total - total number of requests.

These metrics can be filtered by handler, status and request method labels. For example:

codescoring_osa_api_api_http_request_duration_seconds_bucket{handler="/api/osa/packages/",le="0.01",method="POST",status="2xx"} 0.0

Component Status Metrics

• codescoring_osa_api_requested_component_block_status_total is the number of requested components in a particular block status.

This metric can be filtered by the object_type and block_status labels. For example:

codescoring_osa_api_requested_component_block_status_total{block_status="blocked_by_policies",object_type="package"}

• codescoring_osa_api_requested_component_scan_status_total - the number of requested components in a particular scan status.

This metric can be filtered by the object_type and scan_status labels. For example:

codescoring_osa_api_requested_component_scan_status_total{object_type="container_image",scan_status="not_scanned"}