Grafana query for distinct labels number

I have Prometheus counter input_requests for requests with label ticker:

{ticker: TSLA},
{ticker: AAPL},
{ticker: TSLA},
{ticker: META},

I want to do the chart with unique labels number in each period of time, for example 1 minute.

The following PromQL query should return the number of unique values for the label ticker over the last minute:

count(
  count(last_over_time(input_requests[1m])) by (ticker)
)

This query works in the following way:

1. The inner last_over_time(input_requests[1m]) returns the last value per each time series with input_requests name over the last minute. See last_over_time() docs.

2. Then the inner count(...) by (ticker) returns the number of unique time series from the step 1 grouped per each unique ticker value. See count() docs.

3. Then the outer count() returns the number of series returned from the step 2. This is actually the number of unique ticker values seen during the last minute.

Please note that the query above is calculated independently per each point on the graph if the query used in Grafana, e.g. it returns the number of unique ticker values during the last minute ending at every point on the graph. If you want obtaining a graph with per-minute step, then the following options exist:

1. Set the min step in the graph settings to 1m (e.g. one minute).
2. To use Prometheus subqueries in the following way:

last_over_time(
  count(
    count(last_over_time(input_requests[1m] offset -1m)) by (ticker)
  )[1m:1m]
)

It converts the original query to a subquery, which is calculated with 1 minute step, and then the outer last_over_time(...[1m:1m]) returns a graph with 1-minute steps. The query adds negative offset in order to show the number of ticker values seen per the given minute. If the offset isn't set, then the query would show the number of ticker values for the previous minute per each 1-minute step on the graph.

count(changes(input_requests[1m]))