Release v16.31.0

Added

• Persistent event sequence queries in the Workbench: build a query, keep several open as tabs, and organize them in a resizable tree of folders under "Only me" and "Everyone". Which tabs were open, and how wide the tree was left, are restored on the next visit (#1911)
• Filtering on event type, event source, event source type, event stream type, correlation, tags and time range, using the same filter panel the pivot viewer uses. The time range is dragged over a histogram the backend counts across the whole sequence, so it shows the real distribution rather than the loaded page (#1911)
• Sorting the results by clicking a column, and an event sequence picker for targeting any sequence in the store — both saved with the query and applied in storage (#1911)
• Export of the whole filtered result, produced by the server rather than from the page currently on screen (#1911)
• REST endpoints for saving, listing and deleting sequence queries and query folders, and for reading a sequence's events per time bucket
• A paged event sequence query that applies its filter and ordering in storage, so a query over a large sequence only transfers the page being shown
• A Development Tools page for resetting Chronicle, offered only by servers built with development tools compiled in

Changed

• The Sequences page is now the query workspace rather than a single fixed view of the event log. Append, redact, revise and export remain available on the results (#1911)
• Saving a query is a deliberate act instead of happening on every edit: the first save asks for a name, who it is visible to and which folder it belongs in, and later saves write it back where it already lives (#1911)
• A query is renamed where its name is shown — on its node in the tree or by double-clicking its tab — and the new name is saved straight away (#1911)
• Deleting a query or a folder asks for confirmation first, and deleting a folder deletes the queries filed in it (#1911)
• The event sequences an event store offers now come from the EventSequences grain, which reports the sequences a namespace actually holds rather than a hardcoded set
• Event sequence storage can narrow on event source type, event stream type and correlation id, and order by any of the columns a query shows
• Range filters in @cratis/components accept pre-counted histogram buckets, so a range picker can reflect a data set larger than the browser holds

Removed

• The unrouted sequences prototype built on view models over mock data, along with its bookmark tree and mock histogram