Best way to avoid ValueError: Duplicated timeseries in CollectorRegistry

[4815162342lost]

Hello,

I have dynamic metric names (new metrics will be added several times in per day). And if i try to create same double metrics i get an Exception:
ValueError: Duplicated timeseries in CollectorRegistry:

But i do not understand, how can i check if metric already exists. Could you please help me?

For example, code:

promet_metricks = {}
def get_beer_can_name_from_onlime_shop_function():
    ''''In real program this huge function to get beers from online shop and parse available beers'''
    return ("Baltika", "Heinz", "Giguli", "Vodka_lite")

def prom():
    beer_cans = get_beer_can_name_from_onlime_shop_function()
    for current_beer_can in beer_cans:
        promet_metricks[current_beer_can] = prometheus_client.Counter(current_beer_can, "Counter can of beer which drink my dad")
        promet_metricks[current_beer_can].inc()
    time.sleep(3)

while True:
    prom()

How can i check that metric already exists and i should not create them.

Currently i am using

try:
    promet_metricks[current_beer_can] = prometheus_client.Counter(current_beer_can, "Counter can of beer which drink my dad")
except ValueError:
    pass

But i do not think that it is best solution.

[csmarchbanks]

Hello, the best practice is to define the metric name as something static and common. In your example beers_drank_total might be an appropriate name, then use labels to differentiate the types of beer, for example:

beers_drank_total{brand="...", beer="..."}

If your data cannot be modeled in that way I think your solution is as good as any. You will need to keep the metric definitions in state somehow and a dict is a good choice.