I saw the announcement about "Claw Certified" partners and was skeptical from the start. We're evaluating new telemetry pipelines and decided to test three of these certified integrators. The results were embarrassing.
* None could explain the methodology behind Claw's published throughput benchmarks.
* Two proposed architectures that would have broken our existing distributed trace correlation.
* All three pushed Claw's proprietary agents over OpenTelemetry collectors, with zero justification beyond "it's the certified way."
If this is the quality bar for certification, it's a marketing badge, not a technical standard. We're moving forward with an un-certified shop that actually understands the data plane.
/skeptical
Show me the methodology.
Ugh, that's a tough spot. I've seen similar "certified" programs turn into a checkbox exercise for partners who just memorize the vendor's sales deck. The push for proprietary agents over OTel collectors is a major red flag - it's a vendor lock-in play, not an architectural recommendation.
Your point about breaking trace correlation hits home. A while back, a "certified" team nearly deployed a config that would have stripped our `x-b3` headers because they didn't understand the Istio ingress -> internal app -> egress flow. We caught it in a staging cluster review, but it makes you wonder what the certification actually tested.
Did any of them even attempt to show you a comparison manifest or a Helm values diff for using the OTel Collector instead? Or was it just a blanket "use our thing"?
YAML is not a programming language, but I treat it like one.