Would be nice to be able to provided a keen.eventId value and have the event discarded if an event with that ID has already been written to the collection. The main use case for us is if an event write fails, we have no way to know if it was the request that failed or the response. If it was just the response and the event was successfully written retry would lead to a duplicate but not trying again would lead to missing data.

i.e. Support for general purpose indexed use cases. Presenting metrics to your customers typically involves a lot of queries that are "parameterized" or "indexed" by a property such as account_id . Cached datasets provide a mechanism to handle this indexing in many common scenarios, but we think we can do a lot better. We're working on a brand new design to enable you to show much more (and more valuable) information to your customers, without sacrificing speed or cost.

It would be useful to see commas separating numbers into the thousands, millions, etc. It is difficult to see what a number is at first glance, especially once you get past the 1,000s mark

Several customers have requested information regarding the # of events sent to Keen, broken down by event streams, in order to optimize and identify usage per stream.

Currently the API for creating project only allows to specify the S3 bucket where events are streamed ( https://keen.io/docs/api/#create-project ). It doesn't, however, support specifying which type of export to enable. Consider adding a property specifying which type of S3 streaming to enable. This way there will be no manual intervention needed from Keen operations side, when creating new projects.

Refresh once per day for example instead of hourly. In a usecase example, I know that there exist properties that update once a day at a particular time, then I do not need the refresh rate to be set at every 1 hour that passes. If the frequency of update could be set when creating a cached dataset, I could set this to every 24 hours, which would save 23 queries per day. If we have 15 cached datasets the way its setup currently there are (23 15 30) = 10,350 wasted queries per month.

I have multiple projects that represent dev->stage->prod. It would be helpful to be able to copy a dashboard I build on dev into prod.

When retrieving event data on CSV format by specifying an email address, it would be nice to be able to customize the CSV filename.

We've gotten a few requests to support SSO login's for teams. We'd love to hear what providers or specs you'd prefer. Google Business Apps, Github Orgs, SAML2, Auth0 etc.