feat: report panics to Sentry

[eneskaya]

Unfortunately there is some boilerplate required for this to work when it comes to go routines.

[mlabouardy]

this could be refactored to avoid duplicated code. Proposal:

func fetchResources(ctx context.Context, clients []providers.ProviderClient, regions []string, telemetry bool) error {
	for _, client := range clients {
		if client.AWSClient != nil {
			go triggerFetchingWorfklow(ctx, client, "AWS", telemetry, regions, sentry.CurrentHub().Clone())
		} else if client.ScalewayClient != nil {
			go triggerFetchingWorfklow(ctx, client, "Scaleway", telemetry, regions, sentry.CurrentHub().Clone())
		} else if client.MongoDBAtlasClient != nil {
		...
}

func triggerFetchingWorfklow(ctx context.Context, client providers.ProviderClient, provider string, telemetry bool, regions []string, localHub *sentry.Hub) {
	defer func() {
		err := recover()
		log.WithField("err", err).Error(fmt.Sprintf("error fetching %s resources", provider))
		if err != nil {
			localHub.CaptureException(err.(error))
			localHub.Flush(2 * time.Second)
		}
	}()

	localHub.ConfigureScope(func(scope *sentry.Scope) {
		scope.SetTag("provider", provider)
	})

	if telemetry {
		analytics.TrackEvent("fetching_resources", map[string]interface{}{
			"provider": provider,
		})
	}

	switch provider {
	case "AWS":
		aws.FetchResources(ctx, client, regions, db, telemetry, analytics)
	case "Scaleway":
		scaleway.FetchResources(ctx, client, db, telemetry, analytics)
	}
}

[mlabouardy]

All good 🚀