You can use the information_schema.engine_metrics_history view to return information about the resource consumption of your engine. As engines can have 1 or more engine clusters that are transient, this provides visibilty into the utilization of each engine cluster. Each row represents the aggregate resource metrics for a single engine cluster at the given event time. Per default, we collect 2 metric snapshots per minute for the last 30 days. You can use a SELECT query to return the metrics for each engine cluster at a given point in time as shown in the example below.

SELECT
  *
FROM
  information_schema.engine_metrics_history
WHERE
  event_time > now() - INTERVAL '2 hours'
LIMIT 100;

Columns in information_schema.engine_metrics_history

Each row has the following columns with information about each engine cluster.

Column NameData TypeDescription
cluster_ordinalBIGINTthe cluster number
event_timeTIMESTAMPTZtimestamp at which the metrics where collected
cpu_usedDECIMALcurrent CPU utilization (percentage)
memory_usedDECIMALcurrent memory used (percentage)
disk_usedDECIMALcurrently used disk space which encompasses space used for cache and spilling (percentage)
cache_hit_ratioDECIMALcurrent SSD cache hit ratio
spilled_bytesBIGINTamount of spilled data to disk in bytes
running_queriesBIGINTnumber of currently running queries in the system
suspended_queriesBIGINTnumber of queries that have been suspended (not-running and awaiting to be executed)
evicted_bytesBIGINTamount of data evicted from the disk in bytes