When computing bucketed data points, we determine into which bucket each data point goes based on its timestamp. Each bucket specifies a duration and if a data point's timestamp falls within that duration, then it goes into said bucket.
HWKMETRICS-368 we will have support for storing tags with each data point. Suppose we a metric for tracking info about requests for a REST endpoint. We tag each data point with the client id associated with each request to the endpoint. We then want to query and aggregate by client id. If there are 10 different client ids for example, then we would have 10 buckets.
We should reuse the same tag filtering syntax that is used elsewhere for consistency.