sys_metrics: reduce log level when measurement interval is shorter than expected

on highly loaded systems it can happen that the get_metrics() is called
twice within a few houndred milliseconds. Logging a warning in this
case isn't needed, so reduce to info.

on the other hand, 100ms might be to convervative. Patch also
lowers the smallest interval to 10ms

lib/src/system/sys_metrics_processor.cc

@@ -69,9 +69,9 @@ sys_metrics_t sys_metrics_processor::get_metrics()
   uint32_t measure_interval_ms =
       std::chrono::duration_cast<std::chrono::milliseconds>(current_time - last_query_time).count();
 
-  // The time elapsed between 2 measures must be greater that 100 milliseconds.
-  if (measure_interval_ms < 100u) {
-    logger.warning("Interval less than 100ms, skipping measurement.");
+  // The time elapsed between 2 measures must be greater that 10 milliseconds.
+  if (measure_interval_ms < 10u) {
+    logger.info("Interval less than 10ms, skipping measurement.");
     return create_null_metrics();
   }