This message was deleted.
# general
s
This message was deleted.
m
Hey @Matt L Regarding your observation about the source code emitting metrics per data source but not explicitly emitting
***_total
and
***_created
metrics, this behavior is often handled by the Prometheus client library itself, not directly in the application code. When a counter metric is created using the Prometheus client, the library automatically creates both the
***_total
counter and the
***_created
gauge. This is a standard practice in Prometheus client implementations to provide more context and utility for the counters.
1. `***_total`: This suffix is typically used for counter metrics in Prometheus. A counter is a cumulative metric that represents a single monotonically increasing counter whose value can only increase or be reset to zero on restart. For example,
druid_task_pending_count_total
would represent the total number of pending tasks since the start or since the last reset. This is useful for tracking rates of increase over time. 2. `***_created`: This suffix indicates a gauge that represents the creation time of the corresponding
***_total
counter. It's a Unix epoch timestamp (in seconds) showing when the counter was first created (or reset). For instance,
druid_task_pending_count_created
shows the time when the
druid_task_pending_count_total
counter was initialized or reset. This metric can be useful for understanding the age of the counter or for detecting if a counter has been reset.