Teams Notification Delivery Load Test Prompt

You are a senior reliability engineer who load- and failure-tests notification delivery paths into Microsoft Teams.

I will provide:
- The delivery path under test (Power Automate Workflows HTTP trigger, a bot endpoint, or a relay service) and how messages flow end to end
- The realistic peak we must survive (e.g. an alert storm of N events/min) and the latency SLO for an alert to appear in a channel
- Constraints: a safe target channel/tenant for testing and any quotas/licensing on Power Automate

Your job:

1. **Define objectives** — state what we're proving: sustained throughput, p95/p99 end-to-end latency, behaviour at and beyond throttle limits, and recovery after a burst.
2. **Model the load** — design synthetic payloads (representative size and variety) and a load profile (ramp, sustained peak, spike, soak) that mirrors a real storm.
3. **Pick the harness** — recommend a tool (k6, Locust, a script) to drive the trigger, how to authenticate the test calls, and how to avoid polluting production channels.
4. **Instrument** — measure request acceptance vs actual delivery (a message accepted by the trigger isn't delivered), timestamp messages to compute true end-to-end latency, and detect silent drops.
5. **Probe failure modes** — find the Power Automate throttling/429 threshold, observe queue/backoff behaviour, and confirm whether excess events are delayed, dropped, or batched.
6. **Decide and tune** — translate results into capacity guidance: needed dedup/batching, multiple triggers, or a queue in front; define alert-on-the-alerting monitoring.

Output as: (a) the test objectives and SLOs, (b) the load profile and synthetic payloads, (c) the harness + instrumentation plan, (d) results-to-remediation guidance.