Langfuse, Helicone, Honeycomb. What each does well, the schema we built on top, and the three dashboards we check every morning.
Most agent observability tools are LLM wrappers around Datadog. It is not fine for the actual agent operation. Below: what we built with Langfuse, Helicone, and Honeycomb, and the schema that ties them together. The tradeoffs we made and why Every architecture choice in this category is a tradeoff. Here are the ones we have made consciously, and the alternative we did not pick.
Here are the ones we have made consciously, and the alternative we did not pick. We use Pipecat instead of building our own orchestration. The win is months of saved engineering. The cost is being a release behind on a few model integrations. We use Anthropic as the default LLM with OpenAI failover, not the other way around.
Langfuse, Helicone, Honeycomb. What each does well, the schema we built on top, and the three dashboards we check every morning. It is filed under AI Infrastructure because that is where operators looking for this problem actually start, and it is written from production work rather than from a content calendar.