CODEWITHSUNDEEP
duplicatesinfluxdbratethroughputdatadesign
caeus
caeus

Reputation: 3716

how count number of data points, to calculate rate (influxdb)?

So we have a system that takes a lot of queries per minute.

We basically want to know the rate at which our system is processing queries at any given time.

In order to achieve it, we send one point to influx per query. No relevant field, actually.

I'm trying to come up with a way in which I can take a window and count the number of points in such given window, but it just doesn't work.

While writing this I realized I was using the same timestamp and decided to use a different one and NS precision. However, that would ignore repeated entries which can happen (as we receive a lot of queries per minute).

The page suggests that we add a unique tag, but... wouldn't that kill the index?

Upvotes: 1

Views: 356

Answers (1)

phu
phu

Reputation: 103

|> aggregateWindow(every: 5m, fn: count)

gives you the count of points in a 5min window. I found your question while searching for a way to make a graph of time differences between points. It seams when it comes to time calculations in one series, it is a difficult task with influxdb.

Upvotes: 0

Related Questions