Aggregate functions

• https://docs.victoriametrics.com/victoriametrics/metricsql/#aggregate-functions

• https://prometheus.io/docs/prometheus/latest/querying/operators/#aggregation-operators

• Can be used to aggregate the elements of a single instant vector, resulting in a new vector of fewer elements with aggregated values:

• without: removes the listed labels from the result vector

• by: drops labels not listed in the by clause

group(q) by (group_labels)

• Returns 1 per each group_labels for time series returned by q.

# all distinct metrics given a certain label value
group({service="insurance-payments"}) by (__name__)

sum(q) by (group_labels)

• Returns the sum per each group_labels for all the time series returned by q

# returns the rate (bytes/s) of all network interfaces (that was summed up by the sum) for each network interface (the by clausule)
sum(rate(node_network_receive_bytes_total[5m])) by (instance)

sum(http_requests_total) by (application, group)
sum(http_requests_total) without (instance)

max(q) by (group_labels)

• Returns the maximum value per each group_labels for all the time series returned by q

# Is my service up and/or scrapeable?
absent(up{kubernetes_name+"doccserver"}) or
sum(up{kubernetes_name="doccserver"})
== 0

# Do I have the number of LB I expect?
sum(up{kubernetes_name="loadbalancer"}) < 3

# Is out LB at 50% capacity in terms of sessions?
max(haproxy_frontend_current_sessions / haproxy_frontend_limit_sessions)
BY (kubernetes_node_name, frontend) * 100
> 50

# Are 50% of tests taking longer than 10min?
max(test_duration_seconds{quantile="0.5", result="pass"})
BY (test_name)
> 600

count_values(l, v)

• Count number of elements with the same value

count_values("version", build_version)

topk(k, v)

topk(5, http_requests_total)