Ben Keith

> @kweitb Generally the code looks OK but I am not sure it would not be better to use `Sampler` which filter's out span's without a parent. Have you checked...

I think in general the better approach is to have your surrounding Chef recipes invoke the service restart after setting the Python scripts in place, but it isn't clear to...

Yeah that is weird how it refers to `windows_service`...maybe `service` is just an alias for that on Windows or something. Ok, yeah that makes sense about how to handle the...

It is _possible_, but we would rather avoid introducing more "pass down" in the config to make things simpler in the agent. There is a bigger problem with this monitor...

Hi @puremjlee this looks like a network issue on your side. Does your Kubernetes cluster allow outbound traffic from the agent pod? Check what firewalls might be in place. [This...

@dloucasfx have you seen anything like this before with Windows? I suspect it is a firewall or network issue in the environment.

Yes, this is a good approach. The `postgresql` monitor is basically what you are describing. It uses the `sql` monitor code under the covers but provides a bunch of "out...

Yes, it absolutely makes sense. I would make a new monitor `mysql` though and not mess with `collectd/mysql` since it would have nothing to do with collectd.

If you are trying to get the extra metrics for slave status as well as keep the existing collectd/mysql metrics for compatibility, just make a `mysql` monitor that wraps the...

Ok, great. Depending on what you are using to generate the config, you might be able to use [YAML anchors/references](https://yaml.org/spec/1.2/spec.html#id2785586) to deduplicate specific values a bit.