feat: introduce environment-aware JVM resource profiles

[RustamSheoran]

Overview

This PR introduces environment-aware JVM and Kubernetes resource profiles for PH-EE deployments in Mifos Gazelle.

The current deployment configuration relies heavily on static JVM sizing and duplicated service-level heap configuration, despite Gazelle targeting deployment environments with dramatically different runtime characteristics:

• constrained Raspberry Pi / ARM systems

• single-node VM deployments

• larger multi-node Kubernetes clusters

This PR introduces a profile-driven deployment model (micro/std/perf) with container-aware JVM ergonomics and layered Helm configuration composition to improve deployment portability, operational safety, and maintainability.

---

Key Changes

Environment-Aware Deployment Profiles

Introduces layered PH-EE deployment profiles:

• micro

• std

• perf

located under:

config/profiles/phee/

These overlays provide environment-specific Kubernetes resource requests/limits while preserving compatibility with existing deployment flows.

---

Automatic Profile Resolution

Adds automatic profile selection logic based on detected host memory:

Profile | Detected RAM -- | -- micro | < 8GB std | 8GB – 32GB perf | > 32GB

Profile selection remains overrideable and defaults safely when overlays are unavailable.

Implemented in:

src/environmentSetup/helpers.sh

---

Container-Aware JVM Ergonomics

Replaces hardcoded JVM heap sizing (-Xms/-Xmx) with container-aware JVM ergonomics:

-XX:InitialRAMPercentage=75.0
-XX:MaxRAMPercentage=75.0

This allows JVM heap allocation to scale dynamically according to Kubernetes resource limits instead of relying on static service-specific heap sizing.

Additional JVM cleanup includes:

• removal of duplicated hardcoded GC configuration

• removal of duplicated heap sizing fragments

• preservation of existing service-level overrides where required

---

Helm Merge Safety Improvements

During implementation, a critical Helm merge issue was identified:

Helm replaces YAML lists rather than deep-merging them.

Injecting JAVA_TOOL_OPTIONS through profile overlays would have overwritten existing deployment.extraEnvs definitions and silently removed critical configuration values (AWS credentials, S3 configuration, service-specific env vars, etc.).

To avoid this:

• JVM ergonomics are now centralized in base configuration

• overlays only modify map-based resource structures and scalar overrides

• existing environment lists remain fully preserved

This significantly reduces deployment risk and improves long-term maintainability.

---

Files Modified

Added

• config/profiles/phee/micro.yaml

• config/profiles/phee/std.yaml

• config/profiles/phee/perf.yaml

Updated

• config/ph_values.yaml

• src/environmentSetup/helpers.sh

• src/deployer/phee.sh

• src/commandline/commandline.sh

---

Validation

Validated:

• shell syntax (bash -n)

• profile fallback behavior

• Helm overlay merge safety

• backwards compatibility with existing deployment flow

• JVM/container resource alignment

• preservation of existing environment variables

---

Operational Impact

This PR improves:

• deployment portability across hardware classes

• JVM behavior under Kubernetes cgroup limits

• maintainability of PH-EE runtime configuration

• operational consistency across environments

• resource safety for constrained deployments

• deployment ergonomics for future scaling/profiling work

---

Related

Addresses DMP 2026 JVM optimization work / Issue #103.

[tdaly-mifos]

👋 Hi @RustamSheoran — thank you for your pull request.

This PR is currently blocked because we do not have a Contributor License Agreement (CLA) on file for your GitHub account.

To get unblocked:

1. Complete the form at https://mifos.org/about-us/financial-legal/mifos-contributor-agreement
2. Complete the CLA signing process
3. Once verified you will be added to the approved contributors list and this PR check will be cleared