-
Notifications
You must be signed in to change notification settings - Fork 202
Description
In a project we are building, we use RobustConnections to have a reliable way of reestablishing dropped connections. This works great, but sometimes we'd like still like to drop a connection after a certain timeout. For example, if a connection fails to be up after trying to reconnect for 5 minutes, we would like to consider it dead for good and propagate the error condition upwards in our stack. Right now, we use a watchdog task that hooks a RobustConnections close- and reconnect-callbacks to see whether the connection is back up in time, but that feels quite hacky.
I see that #259 adds some kind of global timeout to all operations, but I'm not quite sure if that includes what we'd need (nor do I completely understand the changeset). PR #247 proposes a similar feature, but tries to limit the number of attempts instead of the time taken to establish the connection. Both PRs seem to have stalled; is there a blocker for this where I could help out?
Is there a way this could be implemented easily? I could try and implement the execution timeout and open a pull request, however I do not have the required overview over the library (yet).