Some portion of our user base is using browsers too old to support
toReversed (at least it keeps showing up in docs Sentry errors). The
legacy version is a simple enough change, so let's use that instead.
Should've double-checked this before applying the filter =_=""" but
apparently filtering out `app:///` also filters out legitimate events,
since those files are prefixed by `app:///` as well.
Instead let's keep only `app:///_next` since that should include all our
static JS assets.
* Adding note to broadcast_changes that is uses a private channel.
2 customers confused so far on need for RLS and private flag set to use this function.
* Update apps/docs/content/guides/realtime/subscribing-to-database-changes.mdx
---------
Co-authored-by: Charis <26616127+charislam@users.noreply.github.com>
* improve legend items hover
* show yAxis label
* open range in logs explorer
* add provisioned disk size to Memory Usage chart
* temp new charts
* Network Traffic chart
* sync charts marker
* update api types
* fix layout
* add max memory
* add db size
* chore: remove unused properties
* disk_io_usage
* add max db and pooler connections
* chart header tooltip
* clean charts
* add cpu max
* refactor charts and change max color
* reference line on Database Size chart
* reference line text
* reference lines on Database Size chart
* fix aggregation
* legend spacing
* reference line
* add docsUrl
* chore(docs): turn off sentry sends for local dev
No point wasting quota on local dev, it's easy enough to disable the if
statements if you need to debug something with Sentry.
* ci: Autofix updates from GitHub workflow
* feat: add sign in with solana (web3) configuration UI
* Regenerate the api types.
* fix tiny things
---------
Co-authored-by: Ivan Vasilov <vasilov.ivan@gmail.com>
Couple of DB changes to set up for exposing error codes via the API:
- Add a content schema to keep things organized since we'll be syncing
all content to the DB now. This is exposed via the API so it can be
queried via PostgREST.
- Add tables for tracking error codes.
- Add some utility functions for common tasks.
Don't know why I was using string interpolation instead of GraphQL
variables for the GitHub API calls, leading to all kinds of escaping
problems. Changed to using GraphQL variables properly.
* fix(studio): fallback to count estimate if exact count query timeout
* If the table is partitioned, use estimated count always.
* Minor cleanup.
* fix: add all dependencies to useEffect array
* chore: handle null
---------
Co-authored-by: Ivan Vasilov <vasilov.ivan@gmail.com>
Create index with HNSW fails on prod because it has a version of
pgvector < 0.5.0. Removing for now because the index isn't critical
(we're not using any at the moment and we have few enough rows that it
works fine).
Peter