Behaviour

• Connects to the CDPLogger (or LogServer) holding the data and keeps the connection alive, reconnecting automatically if it drops.
• Replays history fast: while it is behind the newest recorded data it reads in large time spans to catch up, then switches to small spans to track newly logged samples in near real time.
• Delivers a clean stream: duplicate and out-of-order rows are filtered out so KPI math is never corrupted, and quiet periods do not stall processing.
• Reports a failing query for the affected Metric only, so one Metric can flag a problem while the others keep progressing.

Configuration

In normal use the source is added to a MetricsManager through the MetricLoggerDataSource recipe, a LoggerDataSource pre-configured to supply routed CDPLogger historical data and tags for the Metrics defined in the MetricDefinition. The settings are:

• Routing - the routed path of the CDPLogger or LogServer that holds the data to read.
• ReadLagSeconds - how far behind "now" data is read, to let the upstream logger settle the most recent rows before they are used for KPI math.
• Debug - diagnostic verbosity; higher values log more detail about the connection, metadata, and per-sample handling.
• Address, Port, Connected - read-only status of the resolved logger endpoint and the live connection.