Data descriptions¶
The Monitoring daemon publishes three tables.
MonitoringProcesses processes to be monitored; connection and registration status
MonitoringMetrics list of available Monitoring metrics per process
MonitoringState status of each Monitoring metric
All tables are published in a delta format; i.e. only the rows that change are published downstream. On initial connection to the daemon, a process will get a snapshot of the all current state of the tables.
MonitoringProcesses¶
This table contains the list of all processes that the daemon is expecting to connect to and monitor.
Columns:
name name of the process
hostPort expected target host/port to connect to
connected true if a valid handle is established to the process
connectTime non-null if connected is true
registered true if the Refinery process successfully registered with the Daemon
registerTime non-null if registered is true
pid current process ID of the Refinery process
MonitoringMetrics¶
This table contains the list of the available Monitoring metrics from each process, populated when the process registers with the daemon.
Columns:
processName name of the process the metrics relates to
metricName name of the metric
sendType event | timer †
interval if sendType is timer, expected send interval of the metric
stateType currently unused, will always be set to dict
† If sendType is set to timer, interval will also be specified. This allows the daemon to track if a process does not send an update for this metric within the specified interval. If the process does not, the daemon will set metricStale for this metric to true
MonitoringState¶
This table contains a status update of each Monitoring metric.
Columns:
processName name of the process the Monitoring data relates to
metricName name of the metric
generatedTime time the data was generated on the source process that generated it
receiveTime time the data was received by the Monitoring Daemon
metricStale 1b when a timer-based metric does not send an update in the expected time
metric published monitoring data