Hey everyone, I've been setting up a new observability pipeline for our product analytics and have been living in the OpenTelemetry docs lately. I'm trying to move away from a vendor-specific agent, and the OTel Collector seems like the perfect, vendor-neutral middle layer.
I see the `opentelemetry-collector-contrib` chart has been around for a while, but I'm feeling a bit cautious about pulling the trigger. My main concern is stability for production. We're handling a lot of user event data (spans, metrics, logs) and I need reliable ingestion.
For those who have deployed it in a live environment:
* Is the Contrib chart generally considered stable now, or is the "contrib" in the name still a warning sign?
* Specifically, have you run into issues with any of the *receiver* components (like the fluentforward or statsd receivers) being buggy or resource-heavy?
* Any major "gotchas" with the default Helm values? I'm worried about cardinality limits and memory spikes.
I'd love to hear real experiences over sales pitches. We're big on data privacy, so keeping control within our infra is key. Thanks in advance for any insights! 🙏