DatabasesHelm chart

OpenSearch Integration

Monitor cluster health, index metrics, and query performance across your OpenSearch deployments. Get AI-powered anomaly detection and shard allocation intelligence before incidents escalate.

Monitor OpenSearch Book a Demo

Setup

How It Works

Deploy opensearch-exporter via Helm

Add the TigerOps Helm chart to your Kubernetes cluster and point it at your OpenSearch REST endpoint. The exporter uses the _cat, _nodes, _cluster, and _stats APIs — no plugin installation required.

Configure Remote Write

Set your TigerOps remote-write endpoint in the Helm values. Cluster health, shard counts, JVM heap, indexing rates, and search latency metrics begin flowing within minutes.

Enable Index-Level Metrics

Configure index patterns to monitor. TigerOps collects per-index document counts, store size, merge rates, and refresh latency. Wildcard patterns are supported for dynamic index names.

Set Cluster Health Alerts

Define thresholds for cluster status (red/yellow), unassigned shards, and search latency SLOs. TigerOps correlates shard allocation failures with JVM pressure and disk watermark breaches.

Capabilities

What You Get Out of the Box

Cluster Health Monitoring

Real-time cluster status (green/yellow/red), active shards, relocating shards, unassigned shards, and pending tasks. TigerOps alerts on yellow status before it degrades to red.

Index Performance Metrics

Per-index indexing rate, search rate, merge time, refresh latency, flush duration, and store size. Track which indices are driving the most I/O and CPU overhead.

Node-Level Resource Tracking

JVM heap usage and GC pause times per node, OS CPU, disk read/write throughput, and thread pool queue depths for search, indexing, and bulk operations.

Query Latency & Throughput

p50, p95, and p99 search latency per index, query cache hit ratio, fielddata memory usage, and fetch phase timing. Surface slow query patterns before they degrade user experience.

Shard Allocation Intelligence

Monitor shard distribution across nodes, detect hot spots, and track relocation progress. TigerOps identifies nodes with imbalanced shard counts and predicts disk watermark breaches.

AI Anomaly Detection

TigerOps AI baselines your OpenSearch indexing and query patterns, then fires alerts when anomalies deviate from learned norms — catching degradation before threshold-based alerts would trigger.

Configuration

Helm Values for opensearch-exporter

Deploy the TigerOps OpenSearch exporter to your Kubernetes cluster with these Helm values.

opensearch-exporter-values.yaml

# TigerOps OpenSearch exporter Helm values
# helm repo add tigerops https://charts.atatus.net
# helm install opensearch-exporter tigerops/opensearch-exporter -f values.yaml

opensearch:
  # OpenSearch cluster endpoint
  uri: https://opensearch.internal:9200
  # TLS verification
  ssl:
    enabled: true
    ca: /etc/ssl/opensearch/ca.pem

  # Credentials (use fine-grained access control role)
  auth:
    username: tigerops_monitor
    passwordSecret:
      name: opensearch-monitor-secret
      key: password

  # Index patterns to collect per-index metrics
  indices:
    patterns:
      - "logs-*"
      - "app-events-*"
      - "metrics-*"
    # Exclude system indices
    excludePatterns:
      - ".opensearch-*"
      - ".kibana*"

  # Node-level metrics
  nodes:
    enabled: true
    includeIndicesStats: true

  # Cluster-level metrics
  cluster:
    enabled: true
    collectAllIndices: false

remoteWrite:
  endpoint: https://ingest.atatus.net/api/v1/write
  bearerToken: "${TIGEROPS_API_KEY}"
  scrapeInterval: 30s

# Alert thresholds
alerts:
  clusterStatus:
    yellow: warning
    red: critical
  unassignedShards: 1
  searchLatencyP99Ms: 500
  jvmHeapPercent: 85

FAQ

Common Questions

Which OpenSearch versions does TigerOps support?

TigerOps supports OpenSearch 1.x and 2.x, including AWS OpenSearch Service (managed). The exporter uses the public REST API so no version-specific plugins are required. Amazon OpenSearch Serverless is also supported via CloudWatch metric bridging.

What credentials does the monitoring user need?

The monitoring user requires the cluster_monitor and indices_monitor action groups in the OpenSearch security plugin. For AWS OpenSearch Service, a fine-grained access control role with es:ESHttp* permissions on the monitoring paths is sufficient.

How does TigerOps handle dynamic index names?

TigerOps supports glob and regex patterns for index filtering. You can monitor daily indices like logs-* or app-events-* without updating config daily. Cardinality limits prevent metric explosion from high-cardinality index namespaces.

Can TigerOps detect when a node is about to hit a disk watermark?

Yes. TigerOps tracks disk usage per node and computes time-to-watermark based on the current ingestion rate. It fires a predictive alert before OpenSearch enforces the low_watermark (85%) and blocks indexing at the high_watermark (90%).

Does TigerOps integrate with OpenSearch Dashboards?

TigerOps provides a pre-built OpenSearch Dashboard template that is importable as a saved object. It surfaces cluster health, hot nodes, top indices by size and throughput, and search latency trends — all linked to your TigerOps alert timeline.

Get Started

Catch OpenSearch Cluster Issues Before They Go Red

Cluster health monitoring, shard allocation intelligence, and AI anomaly detection. Deploy in 5 minutes.

Start Free Talk to an Engineer