k6runner: handle errors reported by http runners

[roobre]

Previously, k6 runners used to communicate errors by returning a non-200 status code, and not returning the logs of the execution. mq proxy, in addition to returning non-200 status codes, also includes an error code and description for errors that occur while running the script.

With the introduction of https://github.com/grafana/sm-k6-runner/pull/153, we now report as errors exceptions being thrown on the script, such as trying to access undefined variables, or invocations of fail(). In this scenario, users likely expect to see the check reported as a failure, but also to see the logs and metrics produced by it.

This PR changes the error-handling logic, so that the response is always parsed for status code that (should) include it, and the success of the check is calculated by checking if there response contains an error.

[roobre]

Not sure if at this point we should create an error with result.Error and return it 🤔

[The-9880]

It makes sense to do I think

[roobre]

Delved a bit more on what this error is used for, and it seems like it is just log.Warned?

synthetic-monitoring-agent/internal/prober/scripted/scripted.go

Lines 57 to 65 in e3403fb

	func (p Prober) Probe(ctx context.Context, target string, registry *prometheus.Registry, logger logger.Logger) bool {
	success, err := p.script.Run(ctx, registry, logger, p.logger)
	if err != nil {
	p.logger.Warn().Err(err).Msg("running probe")
	return false
	}
	return success
	}

If this is the case we might want to return an error when the code is something that is not-on-us.

[roobre]

Refined (slang for "made more complicated") the error handling logic in the latest commit, LMK what you think :)