Adjust activeConnectionsPerPod for pusher and backend-listen

[thainguyensunya]

Change:

• Adjust activeConnectionsPerPod for pusher and backend-listen for better resource optimization.

[greptile-apps]

Greptile Summary

This PR makes a single-line configuration change in both the pusher and backend-listen production Helm values files, incrementing the activeConnectionsPerPod autoscaling threshold from 20 to 21. This is a marginal (+5%) increase in the HPA target metric, meaning each pod will absorb one additional concurrent connection before the cluster considers spinning up new replicas. With minReplicas: 20 this raises the baseline capacity ceiling from 400 to 420 total connections before scale-out begins, providing a small reduction in unnecessary pod churn.

• Both services are updated symmetrically to activeConnectionsPerPod: 21.
• No other autoscaling parameters (minReplicas, maxReplicas, scale-up/down policies, resource requests/limits) are changed.
• The change is strictly additive and low-risk for a production environment given the conservative scale-down stabilization windows already in place (600 s).

Confidence Score: 5/5

• This PR is safe to merge — it is a minimal configuration tweak with no risk of service disruption.
• The change is confined to a single integer value in two YAML config files. The adjustment is very small (+5%), conservative, and symmetric across both services. The existing autoscaling guardrails (PodDisruptionBudget at 80%, slow scale-down stabilization windows, and unchanged min/max replica counts) fully contain any risk from this change.
• No files require special attention.

Important Files Changed

Filename	Overview
backend/charts/backend-listen/prod_omi_backend_listen_values.yaml	Bumps activeConnectionsPerPod from 20 → 21 in the autoscaling section. The rest of the autoscaling config (minReplicas: 20, maxReplicas: 40, scale-up/down policies) is unchanged. No issues found.
backend/charts/pusher/prod_omi_pusher_values.yaml	Bumps activeConnectionsPerPod from 20 → 21 in the autoscaling section, consistent with the backend-listen change. No issues found.

Flowchart

%%{init: {'theme': 'neutral'}}%%
flowchart TD
    A[HPA polls active connections metric] --> B{activeConnections / numPods > 21?}
    B -- Yes --> C[Scale UP\nstabilizationWindow: 120s\nmax +30% or +5 pods per 60s]
    B -- No --> D{activeConnections / numPods < 21?}
    D -- Yes --> E[Scale DOWN candidate\nstabilizationWindow: 600s\nmax -20% or -1 pod per 120s/300s]
    D -- No --> F[Hold at current replica count\nminReplicas: 20, maxReplicas: 40]
    C --> G[New pod count = ceil total_connections / 21]
    E --> H[minAvailable 80% PodDisruptionBudget enforced]

Loading

_{Last reviewed commit: "Adjust activeConnect..."}

[beastoin]

lgtm @thainguyensunya