log(feedback): Add logging to verify produce to ingest-feedback-events #3450
Conversation
| @@ -599,6 +599,15 @@ impl StoreService { | |||
| event_type = message.variant(), | |||
| topic = topic_name, | |||
| ); | |||
| if topic_name == "ingest-feedback-events" { | |||
| if let KafkaMessage::Event(event_msg) = message { | |||
| relay_log::info!("Producing to ingest-feedback-events.\n start_time={}\n event_id={}\n project_id={}", | |||
There was a problem hiding this comment.
Is there a way to pass a kv dict of "extras" like in python?
There was a problem hiding this comment.
Why are our existing metrics not enough?
We shouldn't log in potentially high throughput code (on info level).
There was a problem hiding this comment.
Ha, nvm, I just read your description. I'll take a look at the metrics tomorrow.
There was a problem hiding this comment.
Sounds good! This log won't be high throughput. Since the topic, consumer, and rollout rate only exist/are set in s4s. And there's little to no feedback sent in s4s, except employees who use the feedback widget on a few sentry pages (ex: performance).
|
I also tried sending feedbacks locally as a sanity check.
At each step of this pipeline, I found logs to verify it's working as expected. I also checked #feed-options-automator to verify the option was set to 1.0 in s4s. So this is my last idea to debug before passing it to ops |
|
I should also mention I made a project in s4s and successfully sent feedback to it. We just have no proof it's going through the new infra |
|
Resolved -- s4s metrics now visible in the dashboards above |
Relates to getsentry/sentry#66100
Temporary debugging for s4s. Right now the consumer is deployed and rollout rate is set only in s4s. We're having trouble finding metrics in datadog and want to log as a sanity check.