Merge pull request #39 from netboxlabs/develop

release 🚚

.github/golangci.yaml

@@ -0,0 +1,40 @@
+run:
+  timeout: 5m
+  modules-download-mode: readonly
+
+output:
+  formats: colored-line-number
+
+linters:
+  enable:
+    - revive
+    - errcheck
+    - unused
+    - staticcheck
+    - ineffassign
+    - govet
+    - gosimple
+    - bodyclose
+    - gci
+    - gofumpt
+
+issues:
+  exclude-use-default: false
+  exclude-rules:
+    - path: /*.go
+      text: "package-comments: should have a package comment"
+      linters:
+        - revive
+
+severity:
+  default-severity: error
+
+linters-settings:
+  gci:
+    sections:
+      - standard
+      - default
+      - prefix(github.com/netboxlabs/orb-agent)
+    custom-order: true
+  go-fumpt:
+    extra-rules: true

.github/workflows/lint.yaml

@@ -0,0 +1,29 @@
+name: Orb Agent - lint
+on:
+  push:
+    branches:
+      - "!release"
+  pull_request:
+
+permissions:
+  contents: read
+
+jobs:
+  golangci:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Go
+        uses: actions/setup-go@v4
+        with:
+          go-version: '1.23'
+          check-latest: true
+      - name: Lint
+        uses: golangci/golangci-lint-action@v3
+        with:
+          version: v1.62
+          working-directory: .
+          args: --config .github/golangci.yaml
+          skip-pkg-cache: true
+          skip-build-cache: true

Makefile

@@ -47,6 +47,14 @@ test-coverage:
 	@cat .coverage/cover.out.tmp | grep -Ev "cmd" > .coverage/cover.out
 	@go tool cover -func=.coverage/cover.out | grep total | awk '{print substr($$3, 1, length($$3)-1)}' > .coverage/coverage.txt
 
+.PHONY: lint
+lint:
+	@golangci-lint run ./... --config .github/golangci.yaml
+
+.PHONY: fix-lint
+fix-lint:
+	@golangci-lint run ./... --config .github/golangci.yaml --fix
+
 agent:
 	docker build --no-cache \
 	  --build-arg GOARCH=$(GOARCH) \

README.md

@@ -2,23 +2,20 @@
 Orb agent is a component of the NetBox Discovery solution. It provides network discovery and observability capabilities and is developed by NetBox Labs.
 
 ## Project status
-
 The Orb agent project is currently in the Public Preview stage. For details, please see [NetBox Labs Product and Feature Lifecycle](https://docs.netboxlabs.com/product_feature_lifecycle/). We actively welcome feedback to help us identify and prioritize bugs, features, and improvements.
 
 ## Getting Started
-To run `orb-agent`, first pull the Docker image from [Docker Hub](https://hub.docker.com/r/netboxlabs/orb-agent):
-
+To get started with `orb-agent`, first pull the Docker image from [Docker Hub](https://hub.docker.com/r/netboxlabs/orb-agent):
 
 ```sh
 docker pull netboxlabs/orb-agent:latest
 ```
 
 ## Orb Agent Configuration
-To run, the Orb agent requires a configuration file. This configuration file consists of three main sections: `Config Manager`, `Backends`, and `Policies`.
-
+The Orb agent requires a configuration file. This file consists of three main sections: `config_manager`, `backends`, and `policies`.
 
 ### Config Manager
-The `Config Manager` section specifies how Orb agent should retrieve it's configuration information. The configuration manager is responsible for processing the configuration to retrieve policies and pass them to the appropriate backend.
+The `config_manager` section specifies how Orb agent should retrieve it's configuration information. The configuration manager is responsible for processing the configuration to retrieve policies and pass them to the appropriate backend.
 
 ```yaml
 orb:
@@ -30,7 +27,7 @@ orb:
 Currently, only the `local` manager is supported, which retrieves policies from the local configuration file passed to the agent.
 
 ### Backends
-The `Backends` section specifies what Orb agent backends should be enabled. Each Orb agent backend offers specific discovery or observability capabilities and may require specific configuration information.  
+The `backends` section specifies what Orb agent backends should be enabled. Each Orb agent backend offers specific discovery or observability capabilities and may require specific configuration information.  
 
 ```yaml
 orb:
@@ -42,38 +39,45 @@ orb:
 Only the `network_discovery` and `device_discovery` backends are currently supported. They do not require any special configuration.
 - [Device Discovery](./docs/backends/device_discovery.md) 
 - [Network Discovery](./docs/backends/network_discovery.md)
-### Commons
-A special `common` subsection under `Backends` defines configuration settings that are shared with all backends. Currently, it supports passing [diode](https://github.com/netboxlabs/diode) server settings to all backends.
+
+#### Common
+A special `common` subsection under `backends` defines configuration settings that are shared with all backends. Currently, it supports passing [diode](https://github.com/netboxlabs/diode) server settings to all backends.
 
 ```yaml
   backends:
       ...
       common:
-      diode:
-        target: grpc://192.168.0.22:8080/diode
-        api_key: ${DIODE_API_KEY}
-        agent_name: agent01
+        diode:
+          target: grpc://192.168.0.22:8080/diode
+          api_key: ${DIODE_API_KEY}
+          agent_name: agent01
 ```
 
-
 ### Policies
-The `Policies` section specifies what discovery policies should be passed to each backend. Policies define specific settings for discovery (such as scheduling and default properties) and the scope (targets). Backends can run multiple policies simultaneously, but for each backend all policies must have a unique name. These policies are defined in the `policies` section and are grouped under a subsection for each backend:
+The `policies` section specifies what discovery policies should be passed to each backend. Policies define specific settings for discovery (such as scheduling and default properties) and the scope (targets). Backends can run multiple policies simultaneously, but for each backend all policies must have a unique name. These policies are defined in the `policies` section and are grouped under a subsection for each backend:
 
  ```yaml
 orb:
   ...
   policies:
     device_discovery:
       device_policy_1:
-        # see device_discovery section
+        # see docs/backends/device_discovery.md
     network_discovery:
       network_policy_1:
-       # see network_discovery section
+       # see docs/backends/network_discovery.md
  ```
 
- ## Configuration samples
-You can find sample configurations [here](./docs/config_samples.md) of how to configure Orb agent to run network and device discoveries.
+## Running the agent
 
-## Required Notice
+To run `orb-agent`, use the following command from the directory where your created your `agent.yaml` file:
 
-Copyright NetBox Labs, Inc.
+```sh
+ docker run -v $(PWD):/opt/orb/ netboxlabs/orb-agent:latest run -c /opt/orb/agent.yaml
+```
+
+### Configuration samples
+You can find complete sample configurations [here](./docs/config_samples.md) of how to configure Orb agent to run network and device discoveries, as well as the relevant `docker run` commands.
+
+## Required Notice
+Copyright NetBox Labs, Inc.

agent/agent.go

@@ -9,7 +9,6 @@ import (
 
 	mqtt "github.com/eclipse/paho.mqtt.golang"
 	"github.com/google/uuid"
-	_ "github.com/mattn/go-sqlite3"
 	"github.com/mitchellh/mapstructure"
 	"github.com/orb-community/orb/fleet"
 	"go.uber.org/zap"
@@ -20,10 +19,9 @@ import (
 	"github.com/netboxlabs/orb-agent/agent/version"
 )
 
-var (
-	ErrMqttConnection = errors.New("failed to connect to a broker")
-)
+const routineKey config.ContextKey = "routine"
 
+// Agent is the interface that all agents must implement
 type Agent interface {
 	Start(ctx context.Context, cancelFunc context.CancelFunc) error
 	Stop(ctx context.Context)
@@ -35,7 +33,7 @@ type orbAgent struct {
 	logger            *zap.Logger
 	config            config.Config
 	client            mqtt.Client
-	agent_id          string
+	agentID           string
 	backends          map[string]backend.Backend
 	backendState      map[string]*backend.State
 	backendsCommon    config.BackendCommons
@@ -49,38 +47,29 @@ type orbAgent struct {
 	heartbeatCancel context.CancelFunc
 
 	// Agent RPC channel, configured from command line
-	baseTopic         string
-	rpcToCoreTopic    string
-	rpcFromCoreTopic  string
-	capabilitiesTopic string
-	heartbeatsTopic   string
-	logTopic          string
+	baseTopic        string
+	rpcFromCoreTopic string
+	heartbeatsTopic  string
 
 	// Retry Mechanism to ensure the Request is received
-	groupRequestTicker     *time.Ticker
 	groupRequestSucceeded  context.CancelFunc
-	policyRequestTicker    *time.Ticker
 	policyRequestSucceeded context.CancelFunc
 
 	// AgentGroup channels sent from core
-	groupsInfos map[string]GroupInfo
+	groupsInfos map[string]groupInfo
 
 	policyManager manager.PolicyManager
-	configManager config.ConfigManager
+	configManager config.Manager
 }
 
-const retryRequestDuration = time.Second
-const retryRequestFixedTime = 15
-const retryDurationIncrPerAttempts = 10
-const retryMaxAttempts = 4
-
-type GroupInfo struct {
+type groupInfo struct {
 	Name      string
 	ChannelID string
 }
 
 var _ Agent = (*orbAgent)(nil)
 
+// New creates a new agent
 func New(logger *zap.Logger, c config.Config) (Agent, error) {
 	pm, err := manager.New(logger, c)
 	if err != nil {
@@ -93,11 +82,10 @@ func New(logger *zap.Logger, c config.Config) (Agent, error) {
 	}
 	cm := config.New(logger, c.OrbAgent.ConfigManager)
 
-	return &orbAgent{logger: logger, config: c, policyManager: pm, configManager: cm, groupsInfos: make(map[string]GroupInfo)}, nil
+	return &orbAgent{logger: logger, config: c, policyManager: pm, configManager: cm, groupsInfos: make(map[string]groupInfo)}, nil
 }
 
 func (a *orbAgent) managePolicies() error {
-
 	if a.config.OrbAgent.Policies == nil {
 		return errors.New("no policies specified")
 	}
@@ -147,7 +135,7 @@ func (a *orbAgent) startBackends(agentCtx context.Context) error {
 			a.logger.Info("failed to configure backend", zap.String("backend", name), zap.Error(err))
 			return err
 		}
-		backendCtx := context.WithValue(agentCtx, "routine", name)
+		backendCtx := context.WithValue(agentCtx, routineKey, name)
 		backendCtx = a.configManager.GetContext(backendCtx)
 		a.backends[name] = be
 		initialState := be.GetInitialState()
@@ -177,12 +165,12 @@ func (a *orbAgent) Start(ctx context.Context, cancelFunc context.CancelFunc) err
 	defer func(t time.Time) {
 		a.logger.Debug("Startup of agent execution duration", zap.String("Start() execution duration", time.Since(t).String()))
 	}(startTime)
-	agentCtx := context.WithValue(ctx, "routine", "agentRoutine")
-	asyncCtx, cancelAllAsync := context.WithCancel(context.WithValue(ctx, "routine", "asyncParent"))
+	agentCtx := context.WithValue(ctx, routineKey, "agentRoutine")
+	asyncCtx, cancelAllAsync := context.WithCancel(context.WithValue(ctx, routineKey, "asyncParent"))
 	a.asyncContext = asyncCtx
 	a.rpcFromCancelFunc = cancelAllAsync
 	a.cancelFunction = cancelFunc
-	a.logger.Info("agent started", zap.String("version", version.GetBuildVersion()), zap.Any("routine", agentCtx.Value("routine")))
+	a.logger.Info("agent started", zap.String("version", version.GetBuildVersion()), zap.Any("routine", agentCtx.Value(routineKey)))
 	mqtt.CRITICAL = &agentLoggerCritical{a: a}
 	mqtt.ERROR = &agentLoggerError{a: a}
 
@@ -212,7 +200,7 @@ func (a *orbAgent) logonWithHeartbeat() {
 }
 
 func (a *orbAgent) logoffWithHeartbeat(ctx context.Context) {
-	a.logger.Debug("stopping heartbeat, going offline status", zap.Any("routine", ctx.Value("routine")))
+	a.logger.Debug("stopping heartbeat, going offline status", zap.Any("routine", ctx.Value(routineKey)))
 	if a.heartbeatCtx != nil {
 		a.heartbeatCancel()
 	}
@@ -222,8 +210,9 @@ func (a *orbAgent) logoffWithHeartbeat(ctx context.Context) {
 		}
 	}
 }
+
 func (a *orbAgent) Stop(ctx context.Context) {
-	a.logger.Info("routine call for stop agent", zap.Any("routine", ctx.Value("routine")))
+	a.logger.Info("routine call for stop agent", zap.Any("routine", ctx.Value(routineKey)))
 	if a.rpcFromCancelFunc != nil {
 		a.rpcFromCancelFunc()
 	}
@@ -256,7 +245,7 @@ func (a *orbAgent) RestartBackend(ctx context.Context, name string, reason strin
 
 	be := a.backends[name]
 	a.logger.Info("restarting backend", zap.String("backend", name), zap.String("reason", reason))
-	a.backendState[name].RestartCount += 1
+	a.backendState[name].RestartCount++
 	a.backendState[name].LastRestartTS = time.Now()
 	a.backendState[name].LastRestartReason = reason
 	a.logger.Info("removing policies", zap.String("backend", name))
@@ -272,7 +261,7 @@ func (a *orbAgent) RestartBackend(ctx context.Context, name string, reason strin
 		a.backendState[name].LastError = fmt.Sprintf("failed to reset backend: %v", err)
 		a.logger.Error("failed to reset backend", zap.String("backend", name), zap.Error(err))
 	}
-	be.SetCommsClient(a.agent_id, &a.client, fmt.Sprintf("%s/?/%s", a.baseTopic, name))
+	be.SetCommsClient(a.agentID, &a.client, fmt.Sprintf("%s/?/%s", a.baseTopic, name))
 
 	return nil
 }
@@ -294,7 +283,7 @@ func (a *orbAgent) RestartAll(ctx context.Context, reason string) error {
 }
 
 func (a *orbAgent) extendContext(routine string) (context.Context, context.CancelFunc) {
-	uuidTraceId := uuid.NewString()
-	a.logger.Debug("creating context for receiving message", zap.String("routine", routine), zap.String("trace-id", uuidTraceId))
-	return context.WithCancel(context.WithValue(context.WithValue(a.asyncContext, "routine", routine), "trace-id", uuidTraceId))
+	uuidTraceID := uuid.NewString()
+	a.logger.Debug("creating context for receiving message", zap.String("routine", routine), zap.String("trace-id", uuidTraceID))
+	return context.WithCancel(context.WithValue(context.WithValue(a.asyncContext, routineKey, routine), config.ContextKey("trace-id"), uuidTraceID))
 }

agent/agent_prof_test.go

@@ -6,7 +6,6 @@ import (
 )
 
 func Test_orbAgent_startBackends(t *testing.T) {
-
 	type args struct {
 		agentCtx context.Context
 	}