Monitoring for FragmentAggregatorModule

[MRiganSUSX]

Upon request from @mroda88, here is an initial version of opmon monitoring for FragmentAggregatorModule.

This includes:

• removing old / obsolete portions of opmon vars
• adding opmon metrics for data requests (received, sent, failed)
• adding opmon metrics for fragments (received, sent, failed)
• additional error checking for the fragment error bits (empty, incomplete, invalid window)
• adding opmon metrics for processing time (avg, min, max) for both data requests and fragments
• associated protobuf messages

The counters for failed are set-up to be cumulative (no resets).

Testing:

• passes unit tests
• passes daqsystemtest_integtest_bundle
• local:
  • Able to see the relevant metrics being published and the counters to change in time:
• local:
  • Able to see the relevant metrics being published and the counters to change in time:

    • Data Requests processing times:

      {
       "time": "2025-06-11T09:48:20.972742681Z",
       "origin": {
        "session": "mrigan-local-test",
        "application": "ru-01",
        "substructure": [
         "fragmentaggregator-ru-01"
        ]
       },
       "custom_origin": {
        "data": "DataRequest"
       },
       "measurement": "dunedaq.dfmodules.opmon.FragmentAggregatorTimeInfo",
       "data": {
        "average_us": {
         "uint8_value": "10"
        },
        "min_us": {
         "uint8_value": "3"
        },
        "max_us": {
         "uint8_value": "218"
        }
       }
      }
      

    • Fragments processing times:

      {
       "time": "2025-06-11T09:48:20.972992235Z",
       "origin": {
        "session": "mrigan-local-test",
        "application": "ru-01",
        "substructure": [
         "fragmentaggregator-ru-01"
        ]
       },
       "custom_origin": {
        "data": "Fragment"
       },
       "measurement": "dunedaq.dfmodules.opmon.FragmentAggregatorTimeInfo",
       "data": {
        "average_us": {
         "uint8_value": "29"
        },
        "max_us": {
         "uint8_value": "169"
        },
        "min_us": {
         "uint8_value": "2"
        }
       }
      }
      

    • Fragments counters:

      {
       "time": "2025-06-11T09:48:20.973457987Z",
       "origin": {
        "session": "mrigan-local-test",
        "application": "ru-01",
        "substructure": [
         "fragmentaggregator-ru-01"
        ]
       },
       "measurement": "dunedaq.dfmodules.opmon.FAFragmentsCounterInfo",
       "data": {
        "fragments_empty": {
         "uint8_value": "4"
        },
        "fragments_failed": {
         "uint8_value": "0"
        },
        "fragments_processed": {
         "uint8_value": "231"
        },
        "fragments_invalid": {
         "uint8_value": "0"
        },
        "fragments_received": {
         "uint8_value": "231"
        },
        "fragments_incomplete": {
         "uint8_value": "0"
        }
       }
      }
      

    • Data requests counters:

      {
       "time": "2025-06-11T09:48:51.078466196Z",
       "origin": {
        "session": "mrigan-local-test",
        "application": "ru-01",
        "substructure": [
         "fragmentaggregator-ru-01"
        ]
       },
       "measurement": "dunedaq.dfmodules.opmon.FADataRequestsCounterInfo",
       "data": {
        "data_requests_failed": {
         "uint8_value": "0"
        },
        "data_requests_processed": {
         "uint8_value": "301"
        },
        "data_requests_received": {
         "uint8_value": "301"
        }
       }
      }
      

[Copilot]

Pull Request Overview

Adds operational monitoring metrics to the FragmentAggregatorModule, including counters for data requests and fragments, error‐bit tracking, timing statistics, and updated protobuf definitions.

• Removes deprecated opmon methods and variables
• Introduces atomic counters and timing logic in header and implementation
• Defines new protobuf messages for counters and timing metrics

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File	Description
schema/dfmodules/opmon/FragmentAggregatorModule.proto	Adds counter and timing message definitions
plugins/FragmentAggregatorModule.hpp	Removes old stats, declares new atomic metrics and generate_opmon_data()
plugins/FragmentAggregatorModule.cpp	Implements metric updates, error‐bit checks, timing, and publishing logic

Comments suppressed due to low confidence (3)

plugins/FragmentAggregatorModule.hpp:95

• [nitpick] The field m_fragments_time_average_us actually accumulates total time, not the average. Renaming it to m_fragments_time_total_us would more accurately reflect its purpose.

std::atomic<metric_counter_type> m_fragments_time_average_us{ 0 };

plugins/FragmentAggregatorModule.cpp:69

• Consider adding unit tests for generate_opmon_data() to validate that counters reset correctly and timing metrics publish expected values under various scenarios.

void
FragmentAggregatorModule::generate_opmon_data()

plugins/FragmentAggregatorModule.cpp:160

• Storing the start timestamp in a shared member risks race conditions if methods run concurrently. Consider using a local variable instead of m_timestamp_before_dr.

m_timestamp_before_dr = get_current_time_us();

[mroda88]

Tested at EHN1 and dashboard are created.