github.com/RedHatInsights/insights-results-aggregator@v1.4.1/docs/prometheus.md (about)

     1  ---
     2  layout: page
     3  nav_order: 8
     4  ---
     5  # Prometheus API
     6  
     7  It is possible to use `/api/v1/metrics` REST API endpoint to read all metrics exposed to Prometheus
     8  or to any tool that is compatible with it.
     9  Currently, the following metrics are exposed:
    10  
    11  1. `consumed_messages` the total number of messages consumed from Kafka
    12  1. `consuming_errors` the total number of errors during consuming messages from Kafka
    13  1. `successful_messages_processing_time` the time to process successfully message
    14  1. `failed_messages_processing_time` the time to process message fail
    15  1. `last_checked_timestamp_lag_minutes` shows how slow we get messages from clusters
    16  1. `produced_messages` the total number of produced messages sent to Payload Tracker's Kafka topic
    17  1. `written_reports` the total number of reports written to the storage
    18  1. `feedback_on_rules` the total number of left feedback
    19  1. `sql_queries_counter` the total number of SQL queries
    20  1. `sql_queries_durations` the SQL queries durations
    21  
    22  Additionally it is possible to consume all metrics provided by Go runtime. There metrics start with
    23  `go_` and `process_` prefixes.
    24  
    25  ## API related metrics
    26  
    27  There are a set of metrics provieded by `insights-operator-utils` library, all
    28  of them related with the API usage. These are the API metrics exposed:
    29  
    30  1. `api_endpoints_requests` the total number of requests per endpoint
    31  1. `api_endpoints_response_time` API endpoints response time
    32  1. `api_endpoints_status_codes` a counter of the HTTP status code responses
    33     returned back by the service
    34  
    35  ## Metrics namespace
    36  
    37  As explained in the [configuration](./configuration) section of this
    38  documentation, a namespace can be provided in order to act as a prefix to the
    39  metric name. If no namespace is provided in the configuration, the metrics will
    40  be exposed as described in this documentation.