Connectors (sources) and OLAP under Connect from References, [Added a lot more example and content to Reference -> Connector YAML] #7652
Conversation
…built off docs-clean-up-main-page
connect/connect/ to connect/sources/ not sure why i exploded in broken links (more than sources, possibly we werent checking certain folders but all green now)
never ends!
|
notes for tomorrow connect/source/connectors to connect/connector/sources write out MotherDuck Doc, and Templating Sources make sure that the links and images are actually working! |
… widths, update home url
created playwrite to click through all links.
makes more sense
|
opening this PR up, let me know your thoughts, at some point i mixed up branches and had to pull the other changes to this one so need to recheck once we merge my other two PRs. This moves a lot of docs around but i ran a custom playwright check for links and returned green for links, but will need to do a final run 👍 |
There was a problem hiding this comment.
- Can we remove a layer of hierarchy here so the list of connectors is 2-layers down, not 3-layers down?
- Can we sort the connectors alphabetically? (Helps for scanning a list for what you're looking for.)
- Optionally, we could have a directory for each "kind" of data source: "Object storage", "Data warehouses", "Databases", "OLAP engines", "Other"
- Should this section be all about connecting data, and we move the Slack connector to the "Integrate" section?
- Can we have either content or a redirect at
docs.rilldata.com/connect:
- Whichever approach we take for item 5, can we, near the very top, describe the different connection strategies (e.g. embedded OLAP + ingest your own data, BYO OLAP, BYO OLAP w/ "native" connectors, etc.)
- These YAML reference links did not take me to the connector-specific section:
Done
Done (above)
I think flat is okay for now :)
It originally lived under Explore/alert/slack kind of hidden so wanted to show that it was a connector but yeah kind of out of place here.
Will add this tidbit of data!
Looking into this why the #... isn't working. |
Kept the tip on motherduck on this page just in case users are directed here.
see comment.
|
Removed Templating from Deploy and moved to Connect! There's a bit more content that I will add in a following PR but wanted to get this one out of door., as its already pretty big. Follow ups are:
|
There was a problem hiding this comment.
- Aesthetically, I had a hard time recognizing which description applied to which connection strategy. It'd help if the description were more "coupled" with the strategy name– the description could be indented at the same level as the bullet point, or the description could even be included in the main bullet point.
- There's something not clear about the distinction between BYO OLAP and BYO OLAP w/ Native Connectors. What about just two categories: Rill-managed OLAP (where options are DuckDB or ClickHouse, depending on scale) and BYO OLAP (where some engines support modeling in the native SQL dialect, some don't).
- There's a lack of parallelism between the
/connect,/connect/data-sources(which doesn't exist :)), and/connect/olappages. I'd expect we have a/connect/data-sourcespage that matches the/connect/olappage, including the connector tiles.
There was a problem hiding this comment.
- With these callouts at a different level-of-indentation as the connection strategies, it's hard to read. Can they be at the same level-of-indentation (to better couple them with the associated strategy), or can we communicate the info in a different way or place?
- It's not consistent that, on the
/connectpage, the sequence is 1. OLAP Engines 2. Data sources; but in the sidebar, the sequence is 1. Data sources 2. OLAP Engines. Let's pick one sequence and be consistent with it. I'm thinking that maybe we should frame this as sequential decisions to make: first decision is which OLAP engine do you want to use; second decision is which data do you want to orchestrate into your OLAP engine.
| - **Dashboards** - Visualize your data with interactive charts and tables | ||
|
|
||
| :::warning OLAP Engine Limitations | ||
| Rill supports connecting your data to both [DuckDB](/connect/olap/duckdb) and [ClickHouse](/connect/olap/clickhouse). However, there are still some features in development for managed ClickHouse. [Contact our team](/contact) for more informaton and scheduled feature releases! |
There was a problem hiding this comment.
It's not obvious how this comment "However, there are still some features in development for managed ClickHouse" relates to this page.
There was a problem hiding this comment.
Maybe there should be a dedicated section to describe how different SQL is needed depending on your OLAP engine (DuckDB or ClickHouse)
There was a problem hiding this comment.
hm, unfortunately its not as simple as just the SQL, we dont support all the source ingestions into CH at this moment...
im wondering if adding a little icon on the button indicating duck/ch support...
There was a problem hiding this comment.
added a small snippet to CH doc on managed clickhouse, and added some missing reminders, depending on how a user might stumble on to our docs
seems to match. OLAP then Data |
additions to CH docs, moved source refresh to build/models added additional button on connect to contact for not listed olap, could be a tip though
There was a problem hiding this comment.
Double-checking that this edit was intentional? This component is used in the "Add Data" modal and now looks like this:
There was a problem hiding this comment.
likely not, will fix
Moving Connector and OLAP to Docs under CONNECT
WIth a lot more UI friendly design! Still working on the actual content
Which means no more duplicate source and olap docs,
Reference (as is explained in the other doc is strictly for YAML parameters and not explanations.
Docs are more wordy and explanatory of what we can connect to as a concept.
Still a bit of work to do but want to open the Draft PR if anyone interested in seeing designs :)