jaeger-kusto

Azure Data Explorer (Kusto) gRPC backend for Jaeger

This is a storage grpc-plugin for Jaeger end-to-end distributed tracing system.

Currently, supports version 1.31.0.

https://www.jaegertracing.io/

Installation and testing

For local testing, you need Docker and docker-compose.

First, you have to have Azure Data Explorer cluster, here's a quickstart: https://docs.microsoft.com/en-us/azure/data-explorer/create-cluster-database-portal

Then create a table:

.create table Spans (
TraceID: string,
SpanID: string,
OperationName: string,
References: dynamic,
Flags: int,
StartTime: datetime,
Duration: timespan,
Tags: dynamic,
Logs: dynamic,
ProcessServiceName: string,
ProcessTags: dynamic,
ProcessID: string
)

Then, you should create json config file:

{
  "clientId": "",
  "clientSecret": "",
  "database": "<database>",
  "endpoint": "https://<cluster>.<region>.kusto.windows.net",
  "tenantId": ""
}

Save this file as jaeger-kusto-config.json in the root of repository.

docker-compose up --build will build the all-in-one Jaeger container and start it together with Hotrod test app.

Jaeger UI will be at http://localhost:16686/, Hotrod test app will be at http://localhost:8080.

You can go to Hotrod test app and generate some spans. They will appear in Kusto approx. in 5 minutes (this can be controlled with IngestionBatching policy). After the spans have been ingested, you will see that UI works.

You can check that jaeger-kusto ingestion is working with this query:

.show commands
| where Database == '<yourdatabase>' and CommandType == 'DataIngestPull'
| top 10 by StartedOn

For production deployment we have these charts: dodopizza/jaeger-kusto-query dodopizza/jaeger-kusto-collector dodopizza/jaeger-kusto-agent. You can view latest tag in docker hub - https://hub.docker.com/r/dodopizza/jaeger-kusto/tags