Skip to content

fix(database_observability): Unregister connection_info metric when DB disconnected #5759

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Closed
matthewnolf wants to merge 1 commit into from
Closed

fix(database_observability): Unregister connection_info metric when DB disconnected #5759

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
77 changes: 67 additions & 10 deletions internal/component/database_observability/mysql/collector/connection_info.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,32 +2,41 @@ package collector

import (
"context"
"database/sql"
"net"
"strings"
"time"

"github.com/go-sql-driver/mysql"
"github.com/grafana/alloy/internal/component/database_observability"
"github.com/prometheus/client_golang/prometheus"
"go.uber.org/atomic"
)

const ConnectionInfoName = "connection_info"
const (
ConnectionInfoName = "connection_info"
connectionInfoInterval = 5 * time.Minute
)

type ConnectionInfoArguments struct {
DB *sql.DB
DSN string
Registry *prometheus.Registry
EngineVersion string
CloudProvider *database_observability.CloudProvider
}

type ConnectionInfo struct {
dbConnection *sql.DB
DSN string
Registry *prometheus.Registry
EngineVersion string
InfoMetric *prometheus.GaugeVec
CloudProvider *database_observability.CloudProvider

running *atomic.Bool
ctx context.Context
cancel context.CancelFunc
}

func NewConnectionInfo(args ConnectionInfoArguments) (*ConnectionInfo, error) {
Expand All @@ -37,9 +46,8 @@ func NewConnectionInfo(args ConnectionInfoArguments) (*ConnectionInfo, error) {
Help: "Information about the connection",
}, []string{"provider_name", "provider_region", "provider_account", "db_instance_identifier", "engine", "engine_version"})

args.Registry.MustRegister(infoMetric)

return &ConnectionInfo{
dbConnection: args.DB,
DSN: args.DSN,
Registry: args.Registry,
EngineVersion: args.EngineVersion,
Expand All @@ -54,6 +62,41 @@ func (c *ConnectionInfo) Name() string {
}

func (c *ConnectionInfo) Start(ctx context.Context) error {
labels, err := c.buildLabels()
if err != nil {
return err
}

c.running.Store(true)
ctx, cancel := context.WithCancel(ctx)
c.ctx = ctx
c.cancel = cancel

c.ping(ctx, labels)

go func() {
defer func() {
c.Registry.Unregister(c.InfoMetric)
c.running.Store(false)
}()

ticker := time.NewTicker(connectionInfoInterval)
defer ticker.Stop()

for {
select {
case <-c.ctx.Done():
return
case <-ticker.C:
c.ping(c.ctx, labels)
}
}
}()

return nil
}

func (c *ConnectionInfo) buildLabels() (prometheus.Labels, error) {
var (
providerName = "unknown"
providerRegion = "unknown"
Expand Down Expand Up @@ -82,9 +125,8 @@ func (c *ConnectionInfo) Start(ctx context.Context) error {
} else {
cfg, err := mysql.ParseDSN(c.DSN)
if err != nil {
return err
return nil, err
}

host, _, err := net.SplitHostPort(cfg.Addr)
if err == nil && host != "" {
if strings.HasSuffix(host, "rds.amazonaws.com") {
Expand All @@ -103,17 +145,32 @@ func (c *ConnectionInfo) Start(ctx context.Context) error {
}
}
}
c.running.Store(true)

c.InfoMetric.WithLabelValues(providerName, providerRegion, providerAccount, dbInstanceIdentifier, engine, c.EngineVersion).Set(1)
return nil
return prometheus.Labels{
"provider_name": providerName,
"provider_region": providerRegion,
"provider_account": providerAccount,
"db_instance_identifier": dbInstanceIdentifier,
"engine": engine,
"engine_version": c.EngineVersion,
}, nil
}

func (c *ConnectionInfo) ping(ctx context.Context, labels prometheus.Labels) {
if err := c.dbConnection.PingContext(ctx); err != nil {
c.Registry.Unregister(c.InfoMetric)
return
}
_ = c.Registry.Register(c.InfoMetric)
c.InfoMetric.With(labels).Set(1)
}

func (c *ConnectionInfo) Stopped() bool {
return !c.running.Load()
}

func (c *ConnectionInfo) Stop() {
c.Registry.Unregister(c.InfoMetric)
c.running.Store(false)
if c.cancel != nil {
c.cancel()
}
}
10 changes: 10 additions & 0 deletions internal/component/database_observability/mysql/collector/connection_info_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import (
"strings"
"testing"

"github.com/DATA-DOG/go-sqlmock"
"github.com/aws/aws-sdk-go-v2/aws/arn"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/testutil"
Expand Down Expand Up @@ -85,9 +86,16 @@ func TestConnectionInfo(t *testing.T) {
}

for _, tc := range testCases {
db, mock, err := sqlmock.New(sqlmock.QueryMatcherOption(sqlmock.QueryMatcherEqual), sqlmock.MonitorPingsOption(true))
require.NoError(t, err)
defer db.Close()

mock.ExpectPing()

reg := prometheus.NewRegistry()

collector, err := NewConnectionInfo(ConnectionInfoArguments{
DB: db,
DSN: tc.dsn,
Registry: reg,
EngineVersion: tc.engineVersion,
Expand All @@ -98,8 +106,10 @@ func TestConnectionInfo(t *testing.T) {

err = collector.Start(t.Context())
require.NoError(t, err)
defer collector.Stop()

err = testutil.GatherAndCompare(reg, strings.NewReader(tc.expectedMetrics))
require.NoError(t, err)
require.NoError(t, mock.ExpectationsWereMet())
}
}
1 change: 1 addition & 0 deletions internal/component/database_observability/mysql/component.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -634,6 +634,7 @@ func (c *Component) startCollectors(serverID string, engineVersion string, parse

// Connection Info collector is always enabled
ciCollector, err := collector.NewConnectionInfo(collector.ConnectionInfoArguments{
DB: c.dbConnection,
DSN: string(c.args.DataSourceName),
Registry: c.registry,
EngineVersion: engineVersion,
Expand Down
75 changes: 66 additions & 9 deletions internal/component/database_observability/postgres/collector/connection_info.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2,33 +2,42 @@ package collector

import (
"context"
"database/sql"
"regexp"
"strings"
"time"

"github.com/grafana/alloy/internal/component/database_observability"
"github.com/prometheus/client_golang/prometheus"
"go.uber.org/atomic"
)

const ConnectionInfoName = "connection_info"
const (
ConnectionInfoName = "connection_info"
connectionInfoInterval = 5 * time.Minute
)

var engineVersionRegex = regexp.MustCompile(`(?P<version>^[1-9]+\.[1-9]+)(?P<suffix>.*)?$`)

type ConnectionInfoArguments struct {
DB *sql.DB
DSN string
Registry *prometheus.Registry
EngineVersion string
CloudProvider *database_observability.CloudProvider
}

type ConnectionInfo struct {
dbConnection *sql.DB
DSN string
Registry *prometheus.Registry
EngineVersion string
InfoMetric *prometheus.GaugeVec
CloudProvider *database_observability.CloudProvider

running *atomic.Bool
ctx context.Context
cancel context.CancelFunc
}

func NewConnectionInfo(args ConnectionInfoArguments) (*ConnectionInfo, error) {
Expand All @@ -38,9 +47,8 @@ func NewConnectionInfo(args ConnectionInfoArguments) (*ConnectionInfo, error) {
Help: "Information about the connection",
}, []string{"provider_name", "provider_region", "provider_account", "db_instance_identifier", "engine", "engine_version"})

args.Registry.MustRegister(infoMetric)

return &ConnectionInfo{
dbConnection: args.DB,
DSN: args.DSN,
Registry: args.Registry,
EngineVersion: args.EngineVersion,
Expand All @@ -55,6 +63,41 @@ func (c *ConnectionInfo) Name() string {
}

func (c *ConnectionInfo) Start(ctx context.Context) error {
labels, err := c.buildLabels()
if err != nil {
return err
}

c.running.Store(true)
ctx, cancel := context.WithCancel(ctx)
c.ctx = ctx
c.cancel = cancel

c.ping(ctx, labels)

go func() {
defer func() {
c.Registry.Unregister(c.InfoMetric)
c.running.Store(false)
}()

ticker := time.NewTicker(connectionInfoInterval)
defer ticker.Stop()

for {
select {
case <-c.ctx.Done():
return
case <-ticker.C:
c.ping(c.ctx, labels)
}
}
}()

return nil
}

func (c *ConnectionInfo) buildLabels() (prometheus.Labels, error) {
var (
providerName = "unknown"
providerRegion = "unknown"
Expand Down Expand Up @@ -84,7 +127,7 @@ func (c *ConnectionInfo) Start(ctx context.Context) error {
} else {
parts, err := ParseURL(c.DSN)
if err != nil {
return err
return nil, err
}
if host, ok := parts["host"]; ok {
if strings.HasSuffix(host, "rds.amazonaws.com") {
Expand All @@ -109,17 +152,31 @@ func (c *ConnectionInfo) Start(ctx context.Context) error {
engineVersion = matches[1]
}

c.running.Store(true)
return prometheus.Labels{
"provider_name": providerName,
"provider_region": providerRegion,
"provider_account": providerAccount,
"db_instance_identifier": dbInstanceIdentifier,
"engine": engine,
"engine_version": engineVersion,
}, nil
}

c.InfoMetric.WithLabelValues(providerName, providerRegion, providerAccount, dbInstanceIdentifier, engine, engineVersion).Set(1)
return nil
func (c *ConnectionInfo) ping(ctx context.Context, labels prometheus.Labels) {
if err := c.dbConnection.PingContext(ctx); err != nil {
c.Registry.Unregister(c.InfoMetric)
return
}
_ = c.Registry.Register(c.InfoMetric)
c.InfoMetric.With(labels).Set(1)
}

func (c *ConnectionInfo) Stopped() bool {
return !c.running.Load()
}

func (c *ConnectionInfo) Stop() {
c.Registry.Unregister(c.InfoMetric)
c.running.Store(false)
if c.cancel != nil {
c.cancel()
}
}
10 changes: 10 additions & 0 deletions internal/component/database_observability/postgres/collector/connection_info_test.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,7 @@ import (
"strings"
"testing"

"github.com/DATA-DOG/go-sqlmock"
"github.com/aws/aws-sdk-go-v2/aws/arn"
"github.com/prometheus/client_golang/prometheus"
"github.com/prometheus/client_golang/prometheus/testutil"
Expand Down Expand Up @@ -87,9 +88,16 @@ func TestConnectionInfo(t *testing.T) {
}

for _, tc := range testCases {
db, mock, err := sqlmock.New(sqlmock.QueryMatcherOption(sqlmock.QueryMatcherEqual), sqlmock.MonitorPingsOption(true))
require.NoError(t, err)
defer db.Close()

mock.ExpectPing()

reg := prometheus.NewRegistry()

collector, err := NewConnectionInfo(ConnectionInfoArguments{
DB: db,
DSN: tc.dsn,
Registry: reg,
EngineVersion: tc.engineVersion,
Expand All @@ -100,8 +108,10 @@ func TestConnectionInfo(t *testing.T) {

err = collector.Start(t.Context())
require.NoError(t, err)
defer collector.Stop()

err = testutil.GatherAndCompare(reg, strings.NewReader(tc.expectedMetrics))
require.NoError(t, err)
require.NoError(t, mock.ExpectationsWereMet())
}
}
1 change: 1 addition & 0 deletions internal/component/database_observability/postgres/component.go
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -508,6 +508,7 @@ func (c *Component) startCollectors(systemID string, engineVersion string, cloud

// Connection Info collector is always enabled
ciCollector, err := collector.NewConnectionInfo(collector.ConnectionInfoArguments{
DB: c.dbConnection,
DSN: string(c.args.DataSourceName),
Registry: c.registry,
EngineVersion: engineVersion,
Expand Down
Loading