158 lines
5.6 KiB
Go
Raw Normal View History

2021-01-03 18:15:44 +05:30
package main
import (
"context"
"flag"
2021-01-03 18:15:44 +05:30
"os"
"time"
2021-01-03 18:15:44 +05:30
"github.com/SigNoz/signoz/pkg/config"
"github.com/SigNoz/signoz/pkg/config/envprovider"
"github.com/SigNoz/signoz/pkg/config/fileprovider"
"github.com/SigNoz/signoz/pkg/query-service/app"
"github.com/SigNoz/signoz/pkg/query-service/constants"
"github.com/SigNoz/signoz/pkg/signoz"
"github.com/SigNoz/signoz/pkg/types/authtypes"
"github.com/SigNoz/signoz/pkg/version"
2021-01-03 18:15:44 +05:30
"go.uber.org/zap"
"go.uber.org/zap/zapcore"
)
func initZapLog() *zap.Logger {
config := zap.NewProductionConfig()
2021-01-03 18:15:44 +05:30
config.EncoderConfig.TimeKey = "timestamp"
config.EncoderConfig.EncodeTime = zapcore.ISO8601TimeEncoder
logger, _ := config.Build()
return logger
}
func main() {
var promConfigPath, skipTopLvlOpsPath string
// disables rule execution but allows change to the rule definition
var disableRules bool
2024-09-12 10:58:07 +05:30
var useLogsNewSchema bool
var useTraceNewSchema bool
// the url used to build link in the alert messages in slack and other systems
var ruleRepoURL, cacheConfigPath, fluxInterval, fluxIntervalForTraceDetail string
var cluster string
var preferSpanMetrics bool
var maxIdleConns int
var maxOpenConns int
var dialTimeout time.Duration
2024-09-12 10:58:07 +05:30
flag.BoolVar(&useLogsNewSchema, "use-logs-new-schema", false, "use logs_v2 schema for logs")
flag.BoolVar(&useTraceNewSchema, "use-trace-new-schema", false, "use new schema for traces")
flag.StringVar(&promConfigPath, "config", "./config/prometheus.yml", "(prometheus config to read metrics)")
flag.StringVar(&skipTopLvlOpsPath, "skip-top-level-ops", "", "(config file to skip top level operations)")
flag.BoolVar(&disableRules, "rules.disable", false, "(disable rule evaluation)")
flag.BoolVar(&preferSpanMetrics, "prefer-span-metrics", false, "(prefer span metrics for service level metrics)")
flag.StringVar(&ruleRepoURL, "rules.repo-url", constants.AlertHelpPage, "(host address used to build rule link in alert messages)")
2023-09-17 10:40:45 +05:30
flag.StringVar(&cacheConfigPath, "experimental.cache-config", "", "(cache config to use)")
flag.StringVar(&fluxInterval, "flux-interval", "5m", "(the interval to exclude data from being cached to avoid incorrect cache for data in motion)")
flag.StringVar(&fluxIntervalForTraceDetail, "flux-interval-trace-detail", "2m", "(the interval to exclude data from being cached to avoid incorrect cache for trace data in motion)")
flag.StringVar(&cluster, "cluster", "cluster", "(cluster name - defaults to 'cluster')")
// Allow using the consistent naming with the signoz collector
flag.StringVar(&cluster, "cluster-name", "cluster", "(cluster name - defaults to 'cluster')")
flag.IntVar(&maxIdleConns, "max-idle-conns", 50, "(number of connections to maintain in the pool, only used with clickhouse if not set in ClickHouseUrl env var DSN.)")
flag.IntVar(&maxOpenConns, "max-open-conns", 100, "(max connections for use at any time, only used with clickhouse if not set in ClickHouseUrl env var DSN.)")
flag.DurationVar(&dialTimeout, "dial-timeout", 5*time.Second, "(the maximum time to establish a connection, only used with clickhouse if not set in ClickHouseUrl env var DSN.)")
flag.Parse()
2021-01-03 18:15:44 +05:30
loggerMgr := initZapLog()
zap.ReplaceGlobals(loggerMgr)
defer loggerMgr.Sync() // flushes buffer, if any
logger := loggerMgr.Sugar()
config, err := signoz.NewConfig(context.Background(), config.ResolverConfig{
Uris: []string{"env:"},
ProviderFactories: []config.ProviderFactory{
envprovider.NewFactory(),
fileprovider.NewFactory(),
},
}, signoz.DeprecatedFlags{
MaxIdleConns: maxIdleConns,
MaxOpenConns: maxOpenConns,
DialTimeout: dialTimeout,
Config: promConfigPath,
})
if err != nil {
zap.L().Fatal("Failed to create config", zap.Error(err))
}
version.Info.PrettyPrint(config.Version)
feat(alertmanager): integrate with ruler (#7222) ### Summary Integrate the new implementations of the alertmanager along with changes to the ruler. This change can be broadly categoried into 3 parts: #### Frontend - The earlier `/api/v1/alerts` api was double encoding the response in json and sending it to the frontend. This PR fixes the json response object. For instance, we have gone from the response `{ "status": "success", "data": "{\"status\":\"success\",\"data\":[{\"labels\":{\"alertname\":\"[platform][consumer] consumer is above 100% memory utilization\",\"bu\":\"platform\",\"...... }` to the response `{"status":"success","data":[{"labels":{"alertname":"[Metrics] Pod CP......` - `msteams` has been changed to `msteamsv2` wherever applicable #### Ruler The following changes have been done in the ruler component: - Removal of the old alertmanager and notifier - The RuleDB methods `Create`, `Edit` and `Delete` have been made transactional - Introduction of a new `testPrepareNotifyFunc` for sending test notifications - Integration with the new alertmanager #### Alertmanager Although a huge chunk of the alertmanagers have been merged in previous PRs (the list can be found at https://github.com/SigNoz/platform-pod/issues/404), this PR takes care of changes needed in order to incorporate it with the ruler - Addition of ruleId based matching - Support for marshalling the global configuration directly from the upstream alertmanager - Addition of orgId to the legacy alertmanager - Support for always adding defaults to both routes and receivers while creating them - Migration to create the required alertmanager tables - Migration for msteams to msteamsv2 has been added. We will start using msteamv2 config for the new alertmanager and keep using msteams for the old one. #### Related Issues / PR's Closes https://github.com/SigNoz/platform-pod/issues/404 Closes https://github.com/SigNoz/platform-pod/issues/176
2025-03-10 01:30:42 +05:30
signoz, err := signoz.New(
context.Background(),
config,
signoz.NewCacheProviderFactories(),
signoz.NewWebProviderFactories(),
signoz.NewSQLStoreProviderFactories(),
signoz.NewTelemetryStoreProviderFactories(),
)
if err != nil {
zap.L().Fatal("Failed to create signoz", zap.Error(err))
}
// Read the jwt secret key
jwtSecret := os.Getenv("SIGNOZ_JWT_SECRET")
if len(jwtSecret) == 0 {
zap.L().Warn("No JWT secret key is specified.")
} else {
zap.L().Info("JWT secret key set successfully.")
}
jwt := authtypes.NewJWT(jwtSecret, 30*time.Minute, 30*24*time.Hour)
2021-01-03 18:15:44 +05:30
serverOptions := &app.ServerOptions{
Config: config,
HTTPHostPort: constants.HTTPHostPort,
PromConfigPath: promConfigPath,
SkipTopLvlOpsPath: skipTopLvlOpsPath,
PreferSpanMetrics: preferSpanMetrics,
PrivateHostPort: constants.PrivateHostPort,
DisableRules: disableRules,
RuleRepoURL: ruleRepoURL,
CacheConfigPath: cacheConfigPath,
FluxInterval: fluxInterval,
FluxIntervalForTraceDetail: fluxIntervalForTraceDetail,
Cluster: cluster,
UseLogsNewSchema: useLogsNewSchema,
UseTraceNewSchema: useTraceNewSchema,
SigNoz: signoz,
Jwt: jwt,
}
2021-01-03 18:15:44 +05:30
server, err := app.NewServer(serverOptions)
if err != nil {
logger.Fatal("Failed to create server", zap.Error(err))
}
if err := server.Start(context.Background()); err != nil {
2021-01-03 18:15:44 +05:30
logger.Fatal("Could not start servers", zap.Error(err))
}
feat(alertmanager): integrate with ruler (#7222) ### Summary Integrate the new implementations of the alertmanager along with changes to the ruler. This change can be broadly categoried into 3 parts: #### Frontend - The earlier `/api/v1/alerts` api was double encoding the response in json and sending it to the frontend. This PR fixes the json response object. For instance, we have gone from the response `{ "status": "success", "data": "{\"status\":\"success\",\"data\":[{\"labels\":{\"alertname\":\"[platform][consumer] consumer is above 100% memory utilization\",\"bu\":\"platform\",\"...... }` to the response `{"status":"success","data":[{"labels":{"alertname":"[Metrics] Pod CP......` - `msteams` has been changed to `msteamsv2` wherever applicable #### Ruler The following changes have been done in the ruler component: - Removal of the old alertmanager and notifier - The RuleDB methods `Create`, `Edit` and `Delete` have been made transactional - Introduction of a new `testPrepareNotifyFunc` for sending test notifications - Integration with the new alertmanager #### Alertmanager Although a huge chunk of the alertmanagers have been merged in previous PRs (the list can be found at https://github.com/SigNoz/platform-pod/issues/404), this PR takes care of changes needed in order to incorporate it with the ruler - Addition of ruleId based matching - Support for marshalling the global configuration directly from the upstream alertmanager - Addition of orgId to the legacy alertmanager - Support for always adding defaults to both routes and receivers while creating them - Migration to create the required alertmanager tables - Migration for msteams to msteamsv2 has been added. We will start using msteamv2 config for the new alertmanager and keep using msteams for the old one. #### Related Issues / PR's Closes https://github.com/SigNoz/platform-pod/issues/404 Closes https://github.com/SigNoz/platform-pod/issues/176
2025-03-10 01:30:42 +05:30
signoz.Start(context.Background())
if err := signoz.Wait(context.Background()); err != nil {
zap.L().Fatal("Failed to start signoz", zap.Error(err))
}
err = server.Stop(context.Background())
feat(alertmanager): integrate with ruler (#7222) ### Summary Integrate the new implementations of the alertmanager along with changes to the ruler. This change can be broadly categoried into 3 parts: #### Frontend - The earlier `/api/v1/alerts` api was double encoding the response in json and sending it to the frontend. This PR fixes the json response object. For instance, we have gone from the response `{ "status": "success", "data": "{\"status\":\"success\",\"data\":[{\"labels\":{\"alertname\":\"[platform][consumer] consumer is above 100% memory utilization\",\"bu\":\"platform\",\"...... }` to the response `{"status":"success","data":[{"labels":{"alertname":"[Metrics] Pod CP......` - `msteams` has been changed to `msteamsv2` wherever applicable #### Ruler The following changes have been done in the ruler component: - Removal of the old alertmanager and notifier - The RuleDB methods `Create`, `Edit` and `Delete` have been made transactional - Introduction of a new `testPrepareNotifyFunc` for sending test notifications - Integration with the new alertmanager #### Alertmanager Although a huge chunk of the alertmanagers have been merged in previous PRs (the list can be found at https://github.com/SigNoz/platform-pod/issues/404), this PR takes care of changes needed in order to incorporate it with the ruler - Addition of ruleId based matching - Support for marshalling the global configuration directly from the upstream alertmanager - Addition of orgId to the legacy alertmanager - Support for always adding defaults to both routes and receivers while creating them - Migration to create the required alertmanager tables - Migration for msteams to msteamsv2 has been added. We will start using msteamv2 config for the new alertmanager and keep using msteams for the old one. #### Related Issues / PR's Closes https://github.com/SigNoz/platform-pod/issues/404 Closes https://github.com/SigNoz/platform-pod/issues/176
2025-03-10 01:30:42 +05:30
if err != nil {
zap.L().Fatal("Failed to stop server", zap.Error(err))
}
err = signoz.Stop(context.Background())
if err != nil {
zap.L().Fatal("Failed to stop signoz", zap.Error(err))
2021-01-03 18:15:44 +05:30
}
}