Skip to main content

Release - October 11th, 2021

· 4 min read

✨ New ✨

Annotations on Graph

Details

Product Component

Metrics Catalog

More Detail

We're excited to announce that annotations are now populated directly on the metric charts! If you create an annotation, it will show up on the graph and will be color-coded based on the annotation impact. You can click through the annotations and modals will pop up to give you details on them.

Verification Steps Confirm your existing annotations are on the graph and that any new annotations show up on the chart for the relevant time period

Cumulative Metrics

Details

Product Component

Metrics Framework

More Details

Transform now supports cumulative metrics in our Metrics Framework. These metrics are useful for measuring values over a certain time window, typically used for WAU and DAU-style metrics (Weekly Active Users and Daily Active Users)

Cumulative metrics aggregate a measure over a given window. If no window is specified, the window is considered infinite and would aggregate all available data.

Verification Steps

Define a cumulative metric using the following syntax:

  # Define the measure and the window.
type: cumulative
type_params:
measures:
- distinct_users
# default window is infinity

Caveats

Cumulative metrics are calculated through a self-join with a date window in a similar manner as below

SELECT 
COUNT(DISTINCT user_id) AS wau
, days.ds
FROM (SELECT DISTINCT ds FROM user_daily) days
JOIN user_daily
ON user_daily.ds BETWEEN DATEADD('day', -7, days.ds) AND days.ds
GROUP BY
days.ds

If the data size is large, this query can be large and slow to execute. To speed up access, we recommend using materializations / fast cache to precompute these results in a daily pipeline.

Fanout Join Support

Details

Product Component

Metrics Framework

More Detail

Previously in the framework, we did not allow joins that would result in fan out. For example, joining a data source with a measure with a foreign key identifier along with a dimension data source with the same foreign key identifier, leading to possibly multiple results for each identifier. However, there are some use cases where this is useful, so the framework now allows these kinds of joins.

Verification Steps

Create a measure data source and a dimension data source with the same foreign identifier, and query for a metric using that measure and a dimension in the aforementioned data source.

Metric Search Live Typeahead

Details

Product Component

Metrics Catalog

More Detail

We've enhanced our metric search so that it includes a live typeahead. When you search, results will autocomplete. If you don't search, you can scroll through all the metrics in your organization.

Verification Steps

Confirm searching for metrics produces suggestions that match your search, and that you can scroll through all metrics.

Tier Updates

Details

Product Component

Metrics Catalog

More Detail

We updated the Metric Tier icons as well as the location. Tier icons are now gold, silver, and bronze to represent 1,2,3 respectively. Additionally, the tier icon on the metric page now exists at the top by the metric name.

Verification Steps

Confirm metric tier icons are now gold, silver, and bronze and that they exist at the top of the metric page.

Dimension Menu Updates

Details

Product Component

More Detail

We've updated the dimension filter/group menu in a few different ways:

  1. The list exposes identifiers and dimensions together. Identifiers have a top heading (in grey), and the corresponding dimensions live underneath them
  2. The Filter values are now in a different menu next to the dimension selection
  3. The dimension menu has search! You can search on either identifiers or metrics.
  4. You can scroll through the dimension menu to load additional dimensions

Verification Steps

Click on Edit Chart and verify the dimension menu and group by menu is updated with the above changes.

🐛 Bug Fixes 🐛

Saved Query not using date preset as chart data date range

Details

Product Component

Metrics Catalog

More Details

We fixed a bug where the saved query date ranges was showing the 30 day window from when the query was saved rather than the last 30 days.

Verification Steps

Verify the save query range accurately reflects the chart presets.