OpenSearch

Monitor OpenSearch cluster health, indexing throughput, search latency, and JVM usage using the prometheus-exporter plugin. Keep your search infrastructure observable alongside the rest of your stack in xScaler.

Pattern: OpenSearch prometheus-exporter plugin → Prometheus scrape → xScaler remote_write

---

Prerequisites

• OpenSearch 1.0 or later
• prometheus-exporter plugin installed on each node
• xScaler tenant credentials (token + tenant ID)

---

Option A — Prometheus Exporter

Install the prometheus-exporter plugin on each OpenSearch node, then restart the service:

bin/opensearch-plugin install prometheus-exporter
# Restart OpenSearch after installation
sudo systemctl restart opensearch

OpenSearch will now expose metrics at :9200/_prometheus/metrics. Configure Prometheus to scrape and remote_write:

scrape_configs:
  - job_name: opensearch
    metrics_path: /_prometheus/metrics
    static_configs:
      - targets: ["localhost:9200"]

remote_write:
  - url: https://euw1-01.m.xscalerlabs.com/api/v1/push
    authorization:
      type: Bearer
      credentials: <token>
    headers:
      X-Scope-OrgID: "<tenant-id>"

Option B — Grafana Alloy

prometheus.scrape "opensearch" {
  targets = [{
    "__address__"      = "localhost:9200",
    "__metrics_path__" = "/_prometheus/metrics",
  }]
  forward_to = [prometheus.remote_write.xscaler.receiver]
}

prometheus.remote_write "xscaler" {
  endpoint {
    url = "https://euw1-01.m.xscalerlabs.com/api/v1/push"
    authorization {
      type        = "Bearer"
      credentials = "<token>"
    }
    headers = { "X-Scope-OrgID" = "<tenant-id>" }
  }
}

Option C — OpenTelemetry Collector

receivers:
  elasticsearch:
    endpoint: http://localhost:9200
    username: admin
    password: <password>
    collection_interval: 30s

processors:
  batch: {}

exporters:
  otlphttp/xscaler:
    endpoint: https://euw1-01.m.xscalerlabs.com
    headers:
      Authorization: "Bearer <token>"
      X-Scope-OrgID: "<tenant-id>"
    compression: gzip

service:
  pipelines:
    metrics:
      receivers: [elasticsearch]
      processors: [batch]
      exporters: [otlphttp/xscaler]

---

Logs

Collect OpenSearch server and slow log files. Add the following to your Alloy config:

local.file_match "opensearch_logs" {
  path_targets = [{
    __address__ = "localhost",
    __path__    = "/var/log/opensearch/*.log",
    instance    = constants.hostname,
    job         = "integrations/opensearch",
  }]
}

loki.source.file "opensearch_logs" {
  targets    = local.file_match.opensearch_logs.targets
  forward_to = [loki.write.xscaler.receiver]
}

loki.write "xscaler" {
  endpoint {
    url = "https://euw1-01.l.xscalerlabs.com/api/v1/logs/push"

    http_client_config {
      authorization {
        type        = "Bearer"
        credentials = env("XSCALER_TOKEN")
      }
    }

    headers = { "X-Scope-OrgID" = env("XSCALER_TENANT_ID") }
  }
}

Key metrics

Metric	Description
opensearch_cluster_status	Cluster health status (0 = green, 1 = yellow, 2 = red)
opensearch_indices_docs_count	Total number of documents across all indices
opensearch_indices_store_size_bytes	Total on-disk size of all indices in bytes
opensearch_jvm_mem_heap_used_percent	JVM heap memory used as a percentage of the configured maximum
opensearch_thread_pool_queue_count	Number of tasks queued in each thread pool
opensearch_search_query_time_seconds	Cumulative time spent executing search queries