azure-loganalytics-exporter

Azure LogAnalytics exporter

Prometheus exporter for Azure LogAnalytics Kusto queries with configurable fields and transformations.

azure-loganalytics-exporter can query configured workspaces or all workspaces in one or multiple subscriptions. The exporter can also cache metrics and servicediscovery information to reduce requests against workspaces and Azure API.

Usage

Usage:
  azure-loganalytics-exporter [OPTIONS]

Application Options:
      --log.debug                     debug mode [$LOG_DEBUG]
      --log.trace                     trace mode [$LOG_TRACE]
      --log.json                      Switch log output to json format [$LOG_JSON]
      --azure.environment=            Azure environment name (default: AZUREPUBLICCLOUD) [$AZURE_ENVIRONMENT]
      --azure.servicediscovery.cache= Duration for caching Azure ServiceDiscovery of workspaces to reduce API calls
                                      (time.Duration) (default: 30m) [$AZURE_SERVICEDISCOVERY_CACHE]
      --loganalytics.workspace=       Loganalytics workspace IDs [$LOGANALYTICS_WORKSPACE]
      --loganalytics.concurrency=     Specifies how many workspaces should be queried concurrently (default: 5)
                                      [$LOGANALYTICS_CONCURRENCY]
  -c, --config=                       Config path [$CONFIG]
      --server.bind=                  Server address (default: :8080) [$SERVER_BIND]
      --server.timeout.read=          Server read timeout (default: 5s) [$SERVER_TIMEOUT_READ]
      --server.timeout.write=         Server write timeout (default: 10s) [$SERVER_TIMEOUT_WRITE]

Help Options:
  -h, --help                          Show this help message

for Azure API authentication (using ENV vars) see https://docs.microsoft.com/en-us/azure/developer/go/azure-sdk-authentication

Configuration file

• see example.yaml

HTTP Endpoints

HINT: parameters of type multiple can be either specified multiple times and/or splits multiple values by comma.

/probe parameters

uses predefined workspace list defined as parameter/environment variable on startup

/probe/workspace parameters

uses dynamically passed workspaces via HTTP query parameter

/probe/subscription parameters

uses Azure service discovery to find all workspaces in one or multiple subscriptions

Global metrics

available on /metrics

AzureTracing metrics

(with 22.2.0 and later)

Azuretracing metrics collects latency and latency from azure-sdk-for-go and creates metrics and is controllable using environment variables (eg. setting buckets, disabling metrics or disable autoreset).

Settings

Examples

see example.yaml for general ingestion metrics (number of rows per second and number of bytes per second per table)

see example.aks-single.yaml for AKS namespace ingestion metrics (number of rows per second and number of bytes per AKS namespace) sending queryies to each LogAnalaytics workspace individually (single mode)

see example.aks-multi.yaml for AKS namespace ingestion metrics (number of rows per second and number of bytes per AKS namespace) sending one query against multiple LogAnalaytics workspaces (multi mode)

more examples of result processing can be found within azure-resourcegraph-expoter (uses same processing library)

Config file:

queries:
  - metric: azure_loganalytics_operationstatus_count
    query: |-
      Operation
      | summarize count() by OperationStatus
    fields:
      - name: count_
        type: value

Metrics:

# HELP azure_loganalytics_operationstatus_count azure_loganalytics_operationstatus_count
# TYPE azure_loganalytics_operationstatus_count gauge
azure_loganalytics_operationstatus_count{OperationStatus="Succeeded",workspaceId="xxxxx-xxxx-xxxx-xxxx-xxxxxxxxx",workspaceTable="PrimaryResult"} 1

Prometheus configuration

predefined workspaces (at startup via parameter/environment variable)

- job_name: azure-loganalytics-exporter
  scrape_interval: 1m
  metrics_path: /probe
  params:
    cache: ["10m"]
    parallel: ["5"]
  static_configs:
  - targets: ["azure-loganalytics-exporter:8080"]

dynamic workspaces (defined in prometheus configuration)

- job_name: azure-loganalytics-exporter
  scrape_interval: 1m
  metrics_path: /probe/workspace
  params:
    workspace:
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
    cache: ["10m"]
    parallel: ["5"]
  static_configs:
  - targets: ["azure-loganalytics-exporter:8080"]

find workspaces with servicediscovery via subscription

- job_name: azure-loganalytics-exporter
  scrape_interval: 1m
  metrics_path: /probe/subscription
  params:
    subscription:
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
      - xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxx
    cache: ["10m"]
    parallel: ["5"]
  static_configs:
  - targets: ["azure-loganalytics-exporter:8080"]