This upgrades mongodb-erlang from v3.0.18 to v3.0.19. This fixes a bug
that occurred when one issued a command specified as a map*. This fix
should not affect present EMQX as this functionality is currently not
used by EMQX. However, I think it is good to do the upgrade anyway in
case we will use the fixed functionality in the future. There is no need
to update the changelog.
* https://github.com/emqx/mongodb-erlang/pull/36
This fixes https://emqx.atlassian.net/browse/EMQX-8648. The issue
described in `EMQX-8648` is that when deleting a non-existing bridge the
server gives a success response. See below:
```
curl --head -u admin:public2 -X 'DELETE' 'http://localhost:18083/api/v5/bridges/webhook:i_do_not_exist'
HTTP/1.1 204 No Content
date: Tue, 03 Jan 2023 16:59:01 GMT
server: Cowboy
```
After the fix, deleting a non existing bridge will give the following
response:
```
HTTP/1.1 404 Not Found
content-length: 49
content-type: application/json
date: Thu, 05 Jan 2023 12:40:35 GMT
server: Cowboy
```
Closes: EMQX-8648
Related issue: https://github.com/emqx/emqx/issues/9683
When the HTTP request for authz fails (e.g.: resource is down or
server is down), then the HTTP authorizer returns `ignore`, which was
not handled correctly by the authorization callback.
This makes the buffer/resource workers always use `replayq` for
queuing, along with collecting multiple requests in a single call.
This is done to avoid long message queues for the buffer workers and
rely on `replayq`'s capabilities of offloading to disk and detecting
overflow.
Also, this deprecates the `enable_batch` and `enable_queue` resource
creation options, as: i) queuing is now always enables; ii) batch_size
> 1 <=> batch_enabled. The corresponding metric
`dropped.queue_not_enabled` is dropped, along with `batching`. The
batching is too ephemeral, especially considering a default batch time
of 20 ms, and is not shown in the dashboard, so it was removed.
To avoid confusion for the users as to what persistence guarantees we
offer when buffering bridges/resources, we will always enable offload
mode for `replayq`. With this, when the buffer size is above the max
segment size, it'll flush the queue to disk, but on recovery after a
restart it'll clean the existing segments rather than resuming from
them.
