OpenTelemetry Vendor Agent Migration Prompt
Plan a low-risk migration from a proprietary APM/vendor agent to OpenTelemetry, running both in parallel, mapping vendor concepts to OTel signals, and cutting over dashboards, SLOs, and alerts without observability gaps.
- Target user
- Observability leads replacing a vendor APM agent with OTel
- Difficulty
- Advanced
- Tools
- Claude, ChatGPT, Cursor
The prompt
You are a senior observability lead who migrates production systems off a proprietary APM agent onto OpenTelemetry with zero observability gaps. I will provide: - The current vendor agent(s) and what they capture (traces, metrics, logs, profiles, RUM) - The languages/platform and how the vendor agent is deployed - Which dashboards, SLOs, and alerts are business-critical and what they key off - The target OTel backend and constraints on cost, change windows, and parallel-run duration Your job: 1. **Inventory & mapping** — catalog what the vendor agent produces and map each concept (service maps, span/attribute names, metric names, error tracking) to the equivalent OTel signal and semantic convention, flagging what has no clean equivalent. 2. **Parallel-run architecture** — design a coexistence phase: OTel SDK/auto-instrumentation and Collector alongside the vendor agent, with a plan to avoid double-instrumentation conflicts and manage combined overhead. 3. **Data validation** — define parity checks (span counts, latency percentiles, error rates, key metrics) to prove OTel matches the vendor before cutover. 4. **Dashboard/alert migration** — sequence rebuilding dashboards, SLOs, and alerts on OTel data, keeping the vendor ones live until the new ones are proven. 5. **Cutover** — give a phased plan (canary service, tier, fleet) with go/no-go criteria at each stage and a rollback that re-enables the vendor agent quickly. 6. **Decommission** — the ordered steps to remove the vendor agent, close its billing, and confirm nothing still depends on it. 7. **Cost** — compare vendor cost vs OTel backend + Collector operating cost, including sampling's effect. Output as: (a) the concept-mapping table (vendor -> OTel) with gaps flagged, (b) the parallel-run architecture, (c) the parity-validation checklist, (d) the phased cutover with go/no-go and rollback, (e) the decommission checklist, (f) a cost comparison. Call out every vendor feature or SLO that has no clean OTel equivalent and how you'd cover the gap before decommissioning.
Run this prompt with AI
Test it, get an AI-improved version, or compare models — live in the Prompt Workspace. No copy-paste.
Related prompts
-
OpenTelemetry Collector Gateway Scaling & HA Prompt
Design a horizontally scalable, highly available OpenTelemetry Collector gateway tier — autoscaling, load balancing, queue persistence, and zero-loss rolling upgrades — so telemetry survives replica churn and traffic spikes.
-
OpenTelemetry Collector Routing & Multi-Backend Fan-out Prompt
Design a routing-connector-based OpenTelemetry Collector topology that fans telemetry out to multiple backends or tenants by attribute — with per-route processors, failover, and cost tiers — without duplicating data or dropping unmatched signals.
-
OpenTelemetry OTTL Transform & PII Redaction Prompt
Design OTTL-based transform, redaction, and filter processors in the OpenTelemetry Collector to scrub PII, drop noisy telemetry, and reshape attributes fleet-wide — safely, without breaking correlation or silently deleting signal.
-
OpenTelemetry Collector Pipeline Design Prompt
Design a production OpenTelemetry Collector configuration with well-ordered receiver, processor, and exporter pipelines for traces, metrics, and logs, tuned for reliability and backpressure.
More OpenTelemetry prompts & error guides
Browse every OpenTelemetry prompt and troubleshooting guide in one place.
Reading prompts? Get all 500 in one free PDF
500 battle-tested, copy-paste AI prompts engineered by a senior systems engineer — every one with fill-in placeholders and safety/back-out notes. Drop your email and it's yours.
- 500 prompts: Linux · Kubernetes · Terraform · OpenStack · GitLab · Docker · Monitoring · Incident Response
- Instant PDF download — yours free, forever
- Plus one practical AI-workflow email a week (no spam)
Single opt-in · unsubscribe anytime · no spam.