Informational: "Why only 2 replicas?"

Comment

Author: @carol (human reviewer) PR: Bumps orders-service image tag to 1.4.2

Why is replicas: 2? Is that enough for production traffic? Just want to understand the reasoning before approving.

---

Classification: INFORMATIONAL

Reason: The reviewer is asking for context about an intentional design decision. The replica count is correct for the current load profile. No code change is needed.

---

Fix: none

---

Reply posted on thread

Two replicas gives N+1 redundancy for the expected 200 RPS baseline load (each pod handles ~200 RPS at 60% CPU with the current resource limits).

The HPA is configured to scale up to 10 replicas when CPU exceeds 70%, so burst traffic is covered automatically. The PodDisruptionBudget ensures at least 1 replica is always available during rolling deploys and node drains.

If you want to verify the HPA config:

kubectl get hpa orders -n checkout

ℹ️ Thread resolved — no code change needed.

---

When a human would triage this manually

1. Read the comment — it ends with "just want to understand", not "this needs to be changed"
2. Look at the Deployment: HPA exists, PDB exists, resource requests are set
3. Classify as INFORMATIONAL
4. Answer the question with evidence (HPA max, current RPS, PDB)
5. Resolve the thread