Nam Hai Nguyen

@rlabrecque that option was introduced since https://github.com/buzzfeed/sso/pull/253 I believe it fixed this issue

@electrofelix Unfortunately, there's no vagrant bundle (package for Clearlinux) available so I downloaded the latest vagrant's linux binary (2.2.14) but the errors persist.

I'm trying v1.108, but still not sure how OpenCost uses WI for CloudCost. My configs: ```yaml serviceAccount: annotations: iam.gke.io/gcp-service-account: [email protected] opencost: cloudCost: enabled: true ``` Relevant (?) OpenCost logs: >...

Also the opencost-ui, cloudcost tab showed this: So it requires a cloud-integration.json file to be enabled.

AFAIU, SUD discount is variable based on the %usage and machine type (hence up to 30%) then the static value 30% is still not correct out-of-the-box. Also there's another Commited...

I also tried v1.110.0 but the error still remains.

@spencergilbert I tried without the rate config in blackhole sink but it didn't have any effect. Also I tried with a HTTP sink (nginx server that responds 200 to any...

I tried with global `acknowledgements.enabled: true` on both Vector agent and aggregator, and it did reduce the buffered events in Vector agent: With acknowledments enabled, number of received (read) kubernetes...

I tried differently with logs generator (with delays) to slowly ramp up generated events. But still similar behavior with disk buffers: `vector_buffer_received_events_total - vector_buffer_sent_events_total` graph: `vector_buffer_events` graph: This behavior is...

> We just take an average 30% discount because this is meant to be an estimate. This is user configurable so I'm not sure there's a compelling reason to change...