This will inevitably fail: it's not generally possible to update
different keys through the same cluster connection, one or more
update will fail with `MOVED` status. This testcase should serve
as a regression test later.
some telemetry events from wolff are discarded:
* dropped:
this is double counted in wolff,
we now only subscribe to the dropped_queue_full event
* retried_failed:
it has different meanings in wolff,
in wolff, it means it's the 2nd (or onward) produce attempt
in EMQX, it means it's eventually failed after some retries
* retried_success
since we are going to handle the success counters in callbac
this having this reported from wolff will only make things
harder to understand
* failed
wolff never fails (unelss drop which is a different counter)
With this, we avoid performing work or replying to callers that are no
longer waiting on a result.
Also introduces two new counters:
- `dropped.expired` :: happens when a request expires before being
sent downstream
- `late_reply` :: when a response is receive from downstream, but the
caller is no longer for a reply because the request has expired, and
the caller might even have retried it.
Related: https://emqx.atlassian.net/browse/EMQX-8692
This should also correctly account for `retried.*` metrics for sync
requests.
Also fixes cases where race conditions for retrying async requests
could potentially lead to inconsistent metrics.
Fixes more cases where a stale reference to `replayq` was being held
accidentally after a `pop`.
the connector schemas are shared between authn, auth and bridges,
the difference is that authn and authz configs are unions like:
[
{mongo_type = rs,
...
}
{another backend config}
]
however the brdige types are maps like, for example:
mongodb_rs.$name {
mongo_type = rs
...
}
in which case, the mongo_type is not required.
in order to keep the schema static as much as possible,
this field is chanegd to 'required => false' with a default value.
However, for authn and authz, the union selector will still raise
exception if the there is no type provided.
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.
