AWS OpenSearch Integration
Monitor cluster health, index metrics, and search latency for managed OpenSearch Service domains. Get AI-powered query spike analysis and JVM pressure alerts before performance degrades.
How It Works
Create IAM Role with CloudWatch Access
Provision an IAM role with cloudwatch:GetMetricData and cloudwatch:ListMetrics permissions scoped to the AWS/ES namespace. TigerOps uses this role for Metric Stream delivery.
Deploy CloudWatch Metric Streams
Run the TigerOps CloudFormation stack to create a Metric Stream for the AWS/ES namespace. All OpenSearch domain metrics start flowing to TigerOps within minutes of deployment.
Tag Your OpenSearch Domains
Apply environment and team tags to your OpenSearch domains. TigerOps uses these tags to group metrics and route alerts to the right on-call team per domain.
Set Search Latency and Shard Alerts
Define SLOs for search latency, indexing latency, and shard count. TigerOps correlates JVM pressure and disk utilization spikes with search performance degradation automatically.
What You Get Out of the Box
Cluster Health Monitoring
Green/yellow/red cluster status tracking, active shards, relocating shards, and initializing shards. Receive instant alerts when cluster status degrades.
Search & Indexing Latency
P50, P90, and P99 search latency and indexing latency per domain. TigerOps correlates latency spikes with JVM heap pressure and segment merge activity.
Index-Level Metrics
Document count, deleted document count, store size, and refresh latency per index. Identify hot indices consuming disproportionate I/O or memory resources.
JVM Heap & GC Pressure
JVM memory pressure percentage, old generation GC count and duration, and young generation GC rates. Alert before heap saturation causes search degradation.
Storage & EBS Metrics
Free storage space, EBS read/write throughput, IOPS utilization, and disk queue depth per node. Prevent out-of-disk incidents with proactive storage trend alerts.
AI Query Spike Analysis
TigerOps AI detects anomalous query rates and correlates them with application deployments, new index mappings, or cron-scheduled bulk operations for root cause clarity.
CloudFormation Stack for OpenSearch Metric Streams
Deploy the TigerOps CloudFormation stack to stream OpenSearch domain metrics in minutes.
# TigerOps CloudFormation — OpenSearch Metric Streams
# aws cloudformation deploy \
# --template-file tigerops-opensearch-streams.yaml \
# --stack-name tigerops-opensearch \
# --capabilities CAPABILITY_IAM
Parameters:
TigerOpsApiKey:
Type: String
NoEcho: true
Resources:
TigerOpsOpenSearchStream:
Type: AWS::CloudWatch::MetricStream
Properties:
Name: tigerops-opensearch-stream
FirehoseArn: !GetAtt TigerOpsDeliveryStream.Arn
RoleArn: !GetAtt MetricStreamRole.Arn
OutputFormat: opentelemetry0.7
IncludeFilters:
- Namespace: AWS/ES
- Namespace: AWS/AOSS
StatisticsConfigurations:
- AdditionalStatistics:
- p50
- p90
- p99
IncludeMetrics:
- Namespace: AWS/ES
MetricName: SearchLatency
- Namespace: AWS/ES
MetricName: IndexingLatency
TigerOpsDeliveryStream:
Type: AWS::KinesisFirehose::DeliveryStream
Properties:
HttpEndpointDestinationConfiguration:
EndpointConfiguration:
Url: https://ingest.atatus.net/api/v1/cloudwatch
AccessKey: !Ref TigerOpsApiKey
RequestConfiguration:
CommonAttributes:
- AttributeName: service
AttributeValue: opensearch
RetryOptions:
DurationInSeconds: 60
# Recommended alert thresholds for OpenSearch:
# JVMMemoryPressure > 85% → Warning
# JVMMemoryPressure > 95% → Critical
# FreeStorageSpace < 20GB → Warning
# ClusterStatus.red > 0 → Critical (immediate)Common Questions
Does TigerOps support both OpenSearch and Elasticsearch domains?
Yes. AWS renamed Elasticsearch Service to OpenSearch Service but both publish metrics to the AWS/ES namespace. TigerOps handles both domain types transparently using the same CloudWatch Metric Stream configuration.
Which CloudWatch metrics does TigerOps collect for OpenSearch?
TigerOps collects all AWS/ES namespace metrics including ClusterStatus, Nodes, SearchableDocuments, FreeStorageSpace, JVMMemoryPressure, CPUUtilization, SearchLatency, IndexingLatency, KibanaHealthyNodes, and more than 40 additional data points.
Can TigerOps alert on OpenSearch index mapping explosions?
Yes. TigerOps tracks field count growth rates and alerts when an index approaches the default 1000-field mapping limit. This catches dynamic mapping issues before they cause cluster-wide performance problems.
How does TigerOps handle multi-AZ OpenSearch deployments?
TigerOps ingests per-node metrics using the NodeId CloudWatch dimension and maps each node to its AZ. You can view per-AZ shard distribution and node-level CPU and JVM metrics on dedicated dashboard panels.
Does TigerOps monitor OpenSearch Serverless?
Yes. OpenSearch Serverless publishes metrics to the AWS/AOSS namespace. TigerOps includes a separate Metric Stream filter for AOSS and provides OCU utilization, indexing rate, and search rate dashboards for serverless collections.
Stop Discovering OpenSearch Degradation After Users Complain
Cluster health monitoring, search latency tracking, and AI query spike analysis. Deploy in 5 minutes.