Closed mimaison closed 4 months ago
@mimaison Do we want to remove all
of the LOG.info
logs and replace the important ones with trace or debug levels? Since the refactor to update the Prometheus version, there doesn't seem to be as many logs.
That said, in the KafkaMetricsCollector
and YammerMetricsCollector
classes, there is more LOG.info
logging that may be unnecssary. For instance:
LOG.info("Yammer metric {} is allowed", prometheusMetricName);
and
LOG.info("labels {}", labels);
and maybe:
LOG.info("metricName group {}, type {}, name {} converted into {}", metricName.getGroup(), metricName.getType(), metricName.getName(), metricNameStr);
Do we need to keep these?
The point of this ticket is to review the current logging and adjust it so it's relevant and not too verbose, especially at INFO level which is usually the default.
I can't comment on each example. For example, while you were working on the reporter which logs did you find useful?
At the moment there is a lot of logging for debugging purposes. We should clean that up and keep important debugging logs with trace or debug levels.