Findings / run 2026-11-02
Datadog: named 9 times, installed 0.
When Claude, GPT and Gemini set up logging and tracing, they installed Grafana/Prometheus (12) and OpenTelemetry (10). Datadog was named in 9 answers and never installed.
| Product | Installed | Claude | GPT | Gemini | Mentioned |
|---|---|---|---|---|---|
| Grafana / Loki / Prometheus | 12 | 4 | 6 | 2 | 21 |
| OpenTelemetry | 10 | 5 | 1 | 4 | 19 |
| Pino / Winston (logs only) | 5 | 1 | 2 | 2 | 8 |
| Sentry performance | 3 | 0 | 1 | 2 | 3 |
| Datadog | 0 | 0 | 0 | 0 | 9 |
| Elastic / ELK | 0 | 0 | 0 | 0 | 4 |
| Honeycomb | 0 | 0 | 0 | 0 | 4 |
| Axiom | 0 | 0 | 0 | 0 | 1 |
| New Relic | 0 | 0 | 0 | 0 | 1 |
Five tasks that add logging, metrics or tracing to an app, twice on each of Claude, GPT and Gemini. The installs went to open tooling: Grafana / Loki / Prometheus in 12 answers, OpenTelemetry in 10, a logger only (Pino, Winston) in 5, Sentry performance in 3.
Datadog was named in 9 answers and installed in none. It shows up as a place the data could go later:
“It exports telemetry in OTLP format, which you can route to Jaeger, Prometheus, Grafana, Datadog, etc.” Gemini, after installing OpenTelemetry
“Log to stdout and let your platform (Docker, Kubernetes, CloudWatch, Datadog, etc.) collect it.” Claude
Honeycomb (4 mentions) and Elastic (4) got the same treatment: zero installs. In error monitoring Datadog was named twice more, and Sentry took all 30.
What it means
Models treat vendor-neutral OpenTelemetry as the safe default and name commercial backends as destinations. For an observability vendor, being the destination isn't nothing, but the agent never runs your agent install or your SDK, so the developer has to come back and choose you on their own.
Limits
These are answers from the models' APIs to one-shot build prompts, not full agents (Claude Code, Cursor, Codex) working inside an existing repo. Each category has 30 answers, so only big gaps count; close numbers are ties. How we measure, and every answer is in the raw data.
See the data
- Logging / observability: all prompts and answers
- Error monitoring: all prompts and answers
- LLM observability: all prompts and answers
More findings
- Supabase is mentioned, never installed: the Postgres host AI agents pick
- Drizzle vs Prisma: which ORM Claude, GPT and Gemini install
- Sentry and Playwright: the only tools AI agents install every time
- Which email API do AI agents install? Resend 16, Postmark 4, Mailgun 0
- Radix UI is named 11 times and never installed: shadcn/ui takes the pick
- Cloudflare R2 vs S3: AI agents name R2, then install the AWS SDK
- pgvector beats Pinecone, Qdrant and Chroma in AI-written RAG code
- Where AI coding agents deploy: Vercel 12, Netlify, Railway, Render and Fly.io 0
Did we get something wrong?
A tool we missed, a pick that looks wrong, a prompt we should add? Tell us. We read every note and fix the data in the next run.
Is your product losing these picks?
Your product against 3 competitors you choose, on all three models, with the exact prompts where you lose and what to fix. $19, delivered within 24 hours.