Add DFaaS plugin implementation and tests

Author: miciav

lb_plugins/plugins/dfaas/README.md

@@ -0,0 +1,161 @@
+# DFaaS Plugin
+
+This plugin implements the legacy DFaaS sampling workflow using k6 on a
+dedicated host and OpenFaaS/Prometheus on the target.
+
+## Architecture
+- Target host: runner + k3s + OpenFaaS + Prometheus + exporters.
+- k6 host: executes load, controlled via SSH from the target host.
+- Network paths:
+  - target -> k6-host over SSH (port 22 by default)
+  - k6-host -> OpenFaaS gateway (port 31112 by default)
+  - target -> Prometheus (NodePort 30411 by default)
+
+## Prerequisites
+- SSH access from target host to k6 host using the configured key.
+- Target host packages: `kubectl`, `helm`, `faas-cli`, `ansible-playbook`.
+- Open ports: 22 (SSH), 31112 (OpenFaaS gateway), 30411 (Prometheus).
+
+## Setup steps
+1. Install k6 on the k6 host:
+   - Run `lb_plugins/plugins/dfaas/ansible/setup_k6.yml` against the k6 host.
+2. Install k3s/OpenFaaS/Prometheus on the target host:
+   - Run `lb_plugins/plugins/dfaas/ansible/setup_target.yml` against the target.
+3. Verify:
+   - `kubectl get nodes` shows Ready.
+   - `faas-cli list --gateway http://<target>:31112` lists functions.
+   - `curl http://<target>:30411/-/ready` returns 200.
+
+## Run flow
+- Generate all function combinations and rate vectors.
+- For each config:
+  - Enforce cooldown (idle CPU/RAM/POWER and replicas < 2).
+  - Generate a k6 script and run it on the k6 host.
+  - Parse the k6 summary and query Prometheus.
+  - Compute overload and skip dominant configs when needed.
+
+## Outputs
+In the workload output directory:
+- `results.csv`, `skipped.csv`, `index.csv` (legacy-compatible headers).
+- `summaries/summary-<config>-iter<iter>-rep<rep>.json`
+- `metrics/metrics-<config>-iter<iter>-rep<rep>.csv`
+- `k6_scripts/config-<config>.js`
+
+## Config loading
+- `config_path` can be passed via workload options to load a YAML/JSON file that
+  contains `common` and `plugins.dfaas` sections.
+- `plugins.dfaas` overrides `common`, and any options passed alongside
+  `config_path` override both.
+
+## Config schema
+Top-level fields (in `plugins.dfaas`):
+- `k6_host` (str): k6 host address.
+- `k6_user` (str): SSH user for the k6 host.
+- `k6_ssh_key` (str): SSH private key path.
+- `k6_port` (int): SSH port.
+- `k6_workspace_root` (str): workspace root on the k6 host.
+- `output_dir` (str): optional output directory override for DFaaS artifacts.
+- `run_id` (str): optional run identifier used for k6 workspace.
+- `gateway_url` (str): OpenFaaS gateway URL.
+- `prometheus_url` (str): Prometheus base URL (default NodePort 30411).
+- `functions` (list): list of function objects (name/method/body/headers/max_rate).
+- `rates` (object): `min_rate`, `max_rate`, `step`.
+- `combinations` (object): `min_functions`, `max_functions` (max is exclusive).
+- `duration` (str): k6 duration string (e.g. `30s`).
+- `iterations` (int): iterations per configuration.
+- `cooldown` (object): `max_wait_seconds`, `sleep_step_seconds`,
+  `idle_threshold_pct`.
+- `overload` (object): `cpu_overload_pct_of_capacity`, `ram_overload_pct`,
+  `success_rate_node_min`, `success_rate_function_min`,
+  `replicas_overload_threshold`.
+- `queries_path` (str): path to the Prometheus queries file.
+- `deploy_functions` (bool): deploy OpenFaaS store functions.
+- `scaphandre_enabled` (bool): enable power metrics via Scaphandre.
+- `function_pid_regexes` (map): optional PID regex per function when Scaphandre is enabled.
+
+Common base fields:
+- `max_retries` (int): retries for the workload (default 0).
+- `timeout_buffer` (int): safety buffer added to expected runtime (default 10).
+- `tags` (list[str]): workload tags.
+
+Function object fields:
+- `name` (str): OpenFaaS function name.
+- `method` (str): HTTP method (GET/POST/etc).
+- `body` (str): request payload (match legacy payloads in
+  `legacy_materials/samples_generator/utils.py`).
+- `headers` (map): HTTP headers.
+- `max_rate` (int, optional): per-function maximum rate (requests/sec).
+
+## Example config (YAML)
+```
+common:
+  timeout_buffer: 10
+
+plugins:
+  dfaas:
+    k6_host: "10.0.0.50"
+    k6_user: "ubuntu"
+    k6_ssh_key: "~/.ssh/id_rsa"
+    k6_port: 22
+
+    gateway_url: "http://<target-ip>:31112"
+    prometheus_url: "http://<target-ip>:30411"
+
+    functions:
+      - name: "figlet"
+        method: "POST"
+        body: "Hello DFaaS!"
+        headers:
+          Content-Type: "text/plain"
+        max_rate: 100
+      - name: "eat-memory"
+        method: "GET"
+        body: ""
+
+    rates:
+      min_rate: 0
+      max_rate: 200
+      step: 10
+
+    combinations:
+      min_functions: 1
+      max_functions: 2
+
+    duration: "30s"
+    iterations: 3
+
+    cooldown:
+      max_wait_seconds: 180
+      sleep_step_seconds: 5
+      idle_threshold_pct: 15
+
+    overload:
+      cpu_overload_pct_of_capacity: 80
+      ram_overload_pct: 90
+      success_rate_node_min: 0.95
+      success_rate_function_min: 0.90
+      replicas_overload_threshold: 15
+
+    queries_path: "lb_plugins/plugins/dfaas/queries.yml"
+    deploy_functions: true
+    scaphandre_enabled: false
+```
+
+## Formal rules (legacy)
+- Rate list: `rates = [min_rate..max_rate step]` inclusive, ascending.
+- Combinations: function sets from `min_functions..max_functions` (max exclusive).
+- Dominance: Config B dominates A if for every function `rate_B >= rate_A` and
+  for at least one function `rate_B > rate_A`. If A is overloaded, skip all
+  dominant configs.
+- Cooldown: wait until CPU/RAM/POWER <= idle + idle * 15% and replicas < 2, with
+  a max wait of 180s.
+- Overload:
+  - Node overloaded if avg success rate < 0.95 OR CPU > 80% capacity OR
+    RAM > 90% OR any function overload.
+  - Function overloaded if success rate < 0.90 OR replicas >= 15.
+
+## Troubleshooting
+- OpenFaaS gateway unreachable: confirm NodePort 31112 and `faas-cli login`.
+- Prometheus query timeouts: verify `prometheus_url` and NodePort 30411.
+- k6 SSH failures: confirm `k6_host`, `k6_user`, and `k6_ssh_key` are correct.
+- Cooldown never completes: check for stuck replicas or sustained CPU/RAM load.

lb_plugins/plugins/dfaas/__init__.py

@@ -0,0 +1 @@
+"""DFaaS workload plugin package."""

lb_plugins/plugins/dfaas/ansible/manifests/prometheus-deployment.yaml

@@ -0,0 +1,42 @@
+apiVersion: apps/v1
+kind: Deployment
+metadata:
+  name: prometheus
+  namespace: openfaas
+  labels:
+    app: prometheus
+spec:
+  replicas: 1
+  selector:
+    matchLabels:
+      app: prometheus
+  template:
+    metadata:
+      labels:
+        app: prometheus
+    spec:
+      securityContext:
+        runAsNonRoot: true
+        runAsUser: 65534
+        fsGroup: 65534
+      serviceAccountName: prometheus
+      containers:
+        - name: prometheus
+          image: prom/prometheus:v2.52.0
+          args:
+            - "--config.file=/etc/prometheus/prometheus.yml"
+            - "--storage.tsdb.path=/prometheus"
+            - "--storage.tsdb.retention.time=24h"
+          ports:
+            - containerPort: 9090
+          volumeMounts:
+            - name: prometheus-config
+              mountPath: /etc/prometheus
+            - name: prometheus-data
+              mountPath: /prometheus
+      volumes:
+        - name: prometheus-config
+          configMap:
+            name: prometheus-config
+        - name: prometheus-data
+          emptyDir: {}

lb_plugins/plugins/dfaas/ansible/manifests/prometheus-serviceaccount.yaml

@@ -0,0 +1,5 @@
+apiVersion: v1
+kind: ServiceAccount
+metadata:
+  name: prometheus
+  namespace: openfaas

lb_plugins/plugins/dfaas/ansible/run_k6.yml

@@ -0,0 +1,61 @@
+---
+- name: Run k6 workload for a single DFaaS config
+  hosts: all
+  become: true
+  vars:
+    k6_workspace_root: "/var/lib/dfaas-k6"
+    target_name: ""
+    run_id: ""
+    config_id: ""
+    script_src: ""
+    summary_dest: ""
+    summary_fetch_dest: ""
+  tasks:
+    - name: Validate required variables
+      ansible.builtin.assert:
+        that:
+          - target_name | length > 0
+          - run_id | length > 0
+          - config_id | length > 0
+          - script_src | length > 0
+        fail_msg: "target_name, run_id, config_id, and script_src are required."
+
+    - name: Build workspace paths
+      ansible.builtin.set_fact:
+        config_dir: "{{ k6_workspace_root }}/{{ target_name }}/{{ run_id }}/{{ config_id }}"
+        script_dest: "{{ k6_workspace_root }}/{{ target_name }}/{{ run_id }}/{{ config_id }}/config-{{ config_id }}.js"
+
+    - name: Ensure config workspace exists
+      ansible.builtin.file:
+        path: "{{ config_dir }}"
+        state: directory
+        mode: "0755"
+
+    - name: Copy k6 script to workspace
+      ansible.builtin.copy:
+        src: "{{ script_src }}"
+        dest: "{{ script_dest }}"
+        mode: "0644"
+
+    - name: Determine summary and log paths
+      ansible.builtin.set_fact:
+        summary_path: >-
+          {{
+            (summary_dest | length > 0)
+            | ternary(summary_dest, config_dir + '/summary.json')
+          }}
+        log_path: "{{ config_dir }}/k6.log"
+
+    - name: Run k6 with summary export
+      ansible.builtin.shell: >
+        k6 run --summary-export "{{ summary_path }}" "{{ script_dest }}"
+        > "{{ log_path }}" 2>&1
+      args:
+        executable: /bin/bash
+
+    - name: Fetch summary JSON to controller
+      ansible.builtin.fetch:
+        src: "{{ summary_path }}"
+        dest: "{{ summary_fetch_dest }}"
+        flat: true
+      when: summary_fetch_dest | length > 0

lb_plugins/plugins/dfaas/ansible/setup_k6.yml

@@ -0,0 +1,82 @@
+---
+- name: Install k6 and prepare workspace
+  hosts: all
+  become: true
+  vars:
+    k6_workspace_root: "/var/lib/dfaas-k6"
+    k6_key_url: "https://dl.k6.io/key.gpg"
+    k6_keyring_path: "/usr/share/keyrings/k6-archive-keyring.gpg"
+    k6_repo: "deb [signed-by={{ k6_keyring_path }}] https://dl.k6.io/deb stable main"
+    k6_version: "0.49.0"
+    k6_arch_map:
+      x86_64: "amd64"
+      amd64: "amd64"
+      aarch64: "arm64"
+      arm64: "arm64"
+    k6_arch: "{{ k6_arch_map.get(ansible_architecture, 'amd64') }}"
+    k6_tarball_url: "https://github.com/grafana/k6/releases/download/v{{ k6_version }}/k6-v{{ k6_version }}-linux-{{ k6_arch }}.tar.gz"
+    k6_extract_dir: "/tmp/k6-v{{ k6_version }}-linux-{{ k6_arch }}"
+  tasks:
+    - name: Install apt dependencies
+      ansible.builtin.apt:
+        name:
+          - ca-certificates
+          - gnupg
+        state: present
+        update_cache: true
+
+    - name: Ensure keyring directory exists
+      ansible.builtin.file:
+        path: "/usr/share/keyrings"
+        state: directory
+        mode: "0755"
+
+    - name: Download k6 repository key
+      ansible.builtin.get_url:
+        url: "{{ k6_key_url }}"
+        dest: "/tmp/k6-key.gpg"
+        mode: "0644"
+
+    - name: Install k6 repository key
+      ansible.builtin.command: "gpg --dearmor -o {{ k6_keyring_path }} /tmp/k6-key.gpg"
+      args:
+        creates: "{{ k6_keyring_path }}"
+
+    - name: Add k6 APT repository
+      ansible.builtin.apt_repository:
+        repo: "{{ k6_repo }}"
+        filename: "k6"
+        state: present
+
+    - name: Install k6 via APT
+      block:
+        - name: Install k6
+          ansible.builtin.apt:
+            name: k6
+            state: present
+            update_cache: true
+      rescue:
+        - name: Download k6 tarball
+          ansible.builtin.get_url:
+            url: "{{ k6_tarball_url }}"
+            dest: "/tmp/k6.tar.gz"
+            mode: "0644"
+
+        - name: Extract k6 tarball
+          ansible.builtin.unarchive:
+            src: "/tmp/k6.tar.gz"
+            dest: "/tmp"
+            remote_src: true
+
+        - name: Install k6 binary
+          ansible.builtin.copy:
+            src: "{{ k6_extract_dir }}/k6"
+            dest: "/usr/local/bin/k6"
+            mode: "0755"
+            remote_src: true
+
+    - name: Ensure k6 workspace root exists
+      ansible.builtin.file:
+        path: "{{ k6_workspace_root }}"
+        state: directory
+        mode: "0755"