Skip to content

Latest commit

 

History

History
98 lines (67 loc) · 4.57 KB

README.md

File metadata and controls

98 lines (67 loc) · 4.57 KB

Azure Data Explorer (Kusto) gRPC backend for Jaeger

This is a storage grpc-plugin for Jaeger end-to-end distributed tracing system and was originally forked from https://github.com/dodopizza/jaeger-kusto and extended now to support OTEL exporter used with ADX.

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, the setup needed for Kusto/ADX exporter with the tables required for storing OTEL traces data can be set up as explained in the documentation here.

The plugin can query OTELTraces table and provide trace UI details on Jaeger

Authentication

Extending the authentication table provided in the Jaeger plugin, the application uses a similar config file to render Jaeger traces as well.

{
  "clientId": "",
  "clientSecret": "",
  "database": "<database>",
  "endpoint": "https://<cluster>.<region>.kusto.windows.net",
  "tenantId": "",
  "traceTableName":"<trace_table>",// defaults to `OTELTraces` if not provided
  "useManagedIdentity": false, // defaults to false, if true, the plugin will use managed identity to authenticate. Use the clientId field to pass the clientId of the managed identity
  "useWorkloadIdentity": false // defaults to false, if true, the plugin will use WorkloadIdentity to authenticate. Note that the plugin will use the default credentials of the VM/Container to authenticate, it will first look for Azure environment variables to authenticate, followed by the workload identity
}

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

Local runs

Plugin can be started as a standalone app (GRPC server):

  • Standalone app (as grpc server). For this mode, use docker compose --file build/server/docker-compose.yml up --build Once this is done, you can run the Jaeger UI on http://localhost:16686 and see the traces in the UI.

Deploying to Kubernetes

The plugin and Jaeger can be deployed to Kubernetes using the provided Helm chart. The Helm chart is available in the build/server/helm folder. The properties can be customized through values.yaml file.

The list of properties that can be customized are:

baseConfig:
  logLevel: 
  logJson: 
  readNoTruncation: 
  readNoTimeout:
authConfig:
  clientId: 
  useManagedIdentity: 
  database: 
  clusterUrl: 
  tenantId: 
  traceTableName: 

table of yaml properties:

Property Description Default
logLevel Log level for the plugin info
logJson Log format false
readNoTruncation In case KustoQueryLimits are hit, use this property to enable no-truncation false
readNoTimeout The default query timeout is 10 minutes which should be sufficient for most cases. In case this needs to be extended to no-timeout false
clientId Client ID for the plugin, represents the ClientId in case of ManagedIdentity. Set it to the AAD APP Id to use AAD Auth ""
clientSecret If AAD Auth is used, set this to the AAD APP Secret for the APP Id ""
tenantId The AAD tenant to use for authentication ""
useManagedIdentity Use managed identity for authentication (Keyless , recommended) false
useWorkloadIdentity Use Azure default credentials (uses workload identity in case it is defined) for authentication false
database Database name to query the traces ""
clusterUrl Cluster URL where the OTEL traces have been ingested ""
traceTableName Trace table name to query "OTELTraces"
image.repository The repository to pull the kusto-jaeger plugin e.g. agramachandran/jaeger-kusto
image.tag The tag of kusto-jaeger-plugin to use e.g. "1.1.0-Preview"
image.pullPolicy Image pull policy "IfNotPresent"

Known Limitations

The plugin is in early development stage (alpha) has the following known limitations:

  • Currently search by tags is not implemented
  • There are deprecated API's in use. These will be fixed in a newer version of the plugin.

Reporting issues

The logging is controlled in the jaeger-kusto-plugin-config.json in the build/server folder. Please change the logLevel to debug to get more detailed logs. This should show the executed query, please execute this query in Kusto and provide the payload as well to debug issues in the applied transformation. Attach both the logs and the payload to troubleshoot the issue.