Use nb_inputs: 2 for Query builder

Queries page loading bar etc.

Add

        display: inline-block;
        vertical-align: middle;
        width: auto;

to rule-filter-container, rule-operator-container > select

For example, I might want to launch a UI that alerts a user if they enter a value that ends with a space (which might be hard for the user to see). Or I might want to alert a user if they enter a commonly mistaken filter value that cannot occur in the data the backend is ingesting.

Remove mentions of query buffer duration in help and others

Will need to add intelligence for determining cardinality of unique lines in time-series mode and error out if it exceeds a number and/or disable the charting.

Or keep a sliding window of the last X windows

A natural way of visualizing data across windows would be to have a timeseries graph across the windows for metrics in each window.

Polar area can just be the same config as Pie/Doughnut for one metric and Radar charts work for multiple metrics but same configuration as Line/Bar charts.

Notably:

Revisit ember-source to 3
ember-tooltips to 3.4+
ember-scroll-to
ember-tether
ember-light-table

https://denali.design/

Features:

Simple shell like access that submits BQL
Autocomplete with Schema knowledge would be nice

Supported in Bullet Core 0.4.3

The Bullet Web Service now returns a TOO_MANY_QUERIES error when there are too many queries in the system and the bullet.max.concurrent.queries setting is exceeded.

The UI should detect this new error and display an appropriate error screen (it currently displays the rate-limit error screen).

The results view would be greatly enhanced (particularly for Distribution queries) if the user could choose columns to pivot on and/or graph them.

This is needed to achieve feature parity with the backend

When settings are specified for different environments, we only merge the defaults one level deep. The 'defaultValues' key in settings especially has a deeply nested structure that would benefit from having some standard values predefined and merged into per-env settings.

Either hitting the rerun button or going back to the query page and changing some parameter (e.g. Top K's K) and running the query, shows the query summary with the old value

Possible cause: the timeSeriesMetric column does not filter out the injected metric keys.

This was intentional when the UI was originally being created and did not have many aggregations. Now, it is useful to have this.

1. Select "Time Based" window
2. Set "Maximum Rows in Window" to blank
3. Switch over to "Record Based" window
4. Run query
5. Get error complaining about blank maximum records

This prevents bad queries from showing up on the queries page with incorrect summaries.

Bullet Core 0.4.3

Right now, it does all unique combinations of the dimensions and one random metric.

Currently fields with the FLOAT datatype are not treated properly in the query builder. For example, when adding a filter for a field that has FLOAT datatype, the only filters available are "is null" and "is not null". FLOAT fields should be treated the same way as LONG and DOUBLE, supporting >=, <=, ==, etc.

Encode a created query into a serialized format, expose an endpoint that can accept a GET request with this serialized format and create a new query. This will let people "share" queries.

Add a button next to edit and rerun to copy to clipboard

When adding a filter, all trailing whitespace in the value entered by the user is currently sent to the backend.

Because of this, users can easily unintentionally send an incorrect filter. For example, a user might try to add a filter "id"="ke8dj21nd", but accidentally send the trailing whitespace to the backend: "id"="ke8dj21nd "

When filters have trailing (or prepending?) whitespace the UI should warn them about this to ensure it is intentional.

This can be done with either a popup? Or maybe an exclamation-mark icon that appears next to the field with a description that appears when a user mouses-over the icon.

LocalStorage is too limiting.

LocalForage seems to be better and there is an Ember adapter for it:
https://github.com/genkgo/ember-localforage-adapter
https://github.com/localForage/localForage