Fixes https://emqx.atlassian.net/browse/EMQX-9129
Currently, if an user configures a bridge with query mode sync, then
all calls to the underlying driver/connector ("inner calls") will
always be synchronous, regardless of its support for async calls.
Since buffer workers always support async queries ("outer calls"), we
should decouple those two call modes (inner and outer), and avoid
exposing the inner call configuration to user to avoid complexity.
There are two situations when we want to force synchronous calls to
the underlying connector even if it supports async:
1) When using `simple_sync_query`, since we are bypassing the buffer
workers;
2) When retrying the inflight window, to avoid overwhelming the
driver.
Those tests in the `flaky` test are really flaky and require lots of
CI retries.
Apparently, the flakiness comes from race conditions from restarting
bridges with the same name too fast between test cases. Previously,
all test cases were sharing the same bridge name (the module name).
Metrics should only be exposed via the /bridges/:id/metrics endpoint,
and not in other operations such as getting the list of all bridges, or
in the response when a bridge has been created. This commit removes all
traces of metrics for the non-dedicated API endpoints.
This was not caught by our tests because we always test bridge types
in isolation. So, if the config only contains ingress-only bridges,
the `on_message_publish` hook is never installed.
In a real system, if there are bridges of mixed types in the config,
the hook might be installed, and `emqx_bridge:get_matched_bridge_id`
would crash when iterating over the ingress bridges.
* Add option to start autocluster. This is useful for scenarios where
a cluster is already running and has some configurations set (via
config handler/cluster rpc) and later another node joins.
* Improve mnesia data directory isolation for nodes. A new dir is set
so that we may avoid intra and inter-suite flakiness due to dirty
tables and schema.
* The janitor is now called synchronously. This ensures the cleanup
is done before the test pid dies.
Kafka Producer and Consumer bridges rely on this prefix for detecting
a dry run and avoid leaking atoms. At some point, this prefix was
changed, effectively disabling the check in Kafka Producer.
1) Make each node have its own isolated data dir to avoid false
negatives.
2) Allow parameterizing the peer module.
3) Fix cluster RPC after a node joins the cluster.
