Troubleshooting a big, complicated, distributed enterprise utility entails challenges like tracing requests throughout a number of providers, figuring out efficiency bottlenecks throughout the stack, and understanding cascading failures between dependent providers. Clients usually must work with remoted knowledge to establish the underlying explanation for the issue. By correlating completely different indicators like logs, traces, metrics, and different efficiency indicators, you will get worthwhile perception into what brought about the issue, the place, and why.
Amazon OpenSearch Service is a managed service to deploy, function, and search knowledge at scale inside AWS. Amazon Managed Grafana is a safe knowledge visualization service to question operational knowledge from a number of sources, together with OpenSearch Service.
On this submit, we present you the way to use these providers to correlate the assorted observability indicators that enhance root trigger evaluation, thereby leading to decreased Imply Time to Decision (MTTR). We additionally present a reference answer that can be utilized at scale for proactive monitoring of enterprise purposes to keep away from an issue earlier than they happen.
Answer overview
The next diagram exhibits the answer structure for accumulating and correlating varied enterprise telemetry indicators at scale.
On the core of this structure are purposes composed of microservices (represented by orange containers) working on Amazon Elastic Kubernetes Service (Amazon EKS). These microservices comprise instrumentation that emit telemetry knowledge within the type of metrics, logs, and traces. This knowledge is exported into the OpenTelemetry Collector, which serves as a central vendor agnostic gateway to gather this knowledge uniformly.
On this submit, we use an OpenTelemetry demo utility as a pattern enterprise utility. Massive enterprise clients usually separate their observability sign knowledge into varied shops for scalability, fault isolation, entry management, and ease of operation. To help in these capabilities, we suggest and use Amazon OpenSearch Ingestion for a serverless, scalable, and absolutely managed knowledge pipeline. We separate log and hint knowledge and ship them to distinct OpenSearch Service domains. The answer additionally sends the metrics knowledge to Amazon Managed Service for Prometheus.
We use Amazon Managed Grafana as an information visualization and analytics platform to question and visualize this knowledge. We additionally present the way to make use of correlations as a worthwhile device to achieve insights from these indicators unfold throughout varied knowledge shops.
The next sections define constructing this structure at scale.
Conditions
Full the next prerequisite steps:
- Provision and configure the Amazon Managed Prometheus workspace to obtain metrics from the OpenTelemetry Collector.
- Create two devoted OpenSearch Service domains (or use present ones) to ingest logs and traces from the OpenTelemetry Collector.
- Create an Amazon Managed Grafana workspace and configure knowledge sources to hook up with Amazon Managed Prometheus and OpenSearch Service.
- Arrange an EKS cluster to deploy purposes and the OpenTelemetry Collector.
Create log and hint OpenSearch Ingestion pipelines
Earlier than organising the ingestion pipelines, you must create the mandatory AWS Identification and Entry Administration (IAM) insurance policies and roles. This course of entails creating two insurance policies for area and OSIS entry, adopted by making a pipeline function that makes use of these insurance policies.
Create a coverage for ingestion
Full the next steps to create an IAM coverage:
- Open the IAM console.
- Select Insurance policies within the navigation pane, then select Create coverage.
- On the JSON tab, enter the next coverage into the editor:
{
"Model": "2012-10-17",
"Assertion": [
{
"Effect": "Allow",
"Action": "es:DescribeDomain",
"Resource": "arn:aws:es:*:{accountId}:domain/*"
},
{
"Effect": "Allow",
"Action": [ "es:ESHttpGet", "es:HttpHead", "es:HttpDelete", "es:HttpPatch", "es:HttpPost", "es:HttpPut" ],
"Useful resource": "arn:aws:es:us-east-1:{accountId}:area/otel-traces"
},
{
"Impact": "Permit",
"Motion": [ "es:ESHttpGet", "es:HttpHead", "es:HttpDelete", "es:HttpPatch", "es:HttpPost", "es:HttpPut" ],
"Useful resource": "arn:aws:es:us-east-1:{accountId}:area/otel-logs"
}
}
]
}
// Substitute {accountId} with your individual values- Select Subsequent, select Subsequent once more, and title your coverage domain-policy.
- Select Create coverage.
- Create one other coverage with the title osis-policy and use the next JSON:
{
"Model": "2012-10-17",
"Assertion": [
{
"Effect": "Allow",
"Action": "osis:Ingest",
"Resource": "arn:aws:osis:us-east-1:{accountId}:pipeline/osi-pipeline-otellogs"
},
{
"Effect": "Allow",
"Action": "osis:Ingest",
"Resource": "arn:aws:osis:us-east-1:{accountId}:pipeline/osi-pipeline-oteltraces"
}
]
}
// Substitute {accountId} with your individual valuesCreate a pipeline function
Full the next steps to create a pipeline function:
- On the IAM console, select Roles within the navigation pane, then select Create function.
- Choose Customized belief coverage and enter the next coverage into the editor:
{
"Model": "2012-10-17",
"Assertion": [
{
"Effect": "Allow",
"Principal": {
"Service": [
"eks.amazonaws.com",
"osis-pipelines.amazonaws.com"
],
"AWS": "{nodegroup_arn}"
},
"Motion": "sts:AssumeRole"
}
]
}
// Substitute {nodegroup_arn} with your individual values- Select Subsequent, then seek for and choose the insurance policies
osis-policyanddomain-policyyou simply created. - Select Subsequent and title the function
PipelineRole. - Select Create function.
Permit entry for the pipeline function in OpenSearch Service domains
To allow entry for the pipeline function in OpenSearch Service domains, full the next steps:
- Open the OpenSearch Service console.
- Select your area (both logs or traces).
- Select the OpenSearch Dashboards URL
- Check in together with your credentials.
Then, full the next steps for every OpenSearch Service area (logs and traces domains).
- In OpenSearch Dashboards, go to the Safety
- Select Roles after which all_access.
This process makes use of the all_access function for demonstration functions solely. This grants full administrative privileges to the pipeline function, which violates the precept of least privilege and will pose safety dangers. For manufacturing environments, you need to create a customized function with minimal permissions required for knowledge ingestion, restrict permissions to particular indexes and operations, contemplate implementing index patterns and time-based entry controls, and recurrently audit function mappings and permissions. For detailed steerage on creating customized roles with acceptable permissions, seek advice from Safety in Amazon OpenSearch Service.
- Select Mapped customers after which Managed mapping.
- On the Map consumer web page, underneath Backend roles, replace the backend function with the Amazon Useful resource Title (ARN) for the function
PiplelineRole. - Select Map.
Create a pipeline for logs
Full the next steps to create a pipeline for logs:
- Open the OpenSearch Service console.
- Select Ingestion pipelines.
- Select Create pipeline.
- Outline the pipeline configuration by getting into the next:
model: "2"
otel-logs-pipeline:
supply:
otel_logs_source:
path: "/v1/logs"
sink:
- opensearch:
hosts: ["{OpenSearch_domain_endpoint}"]
aws:
sts_role_arn: "arn:aws:iam::{accountId}:function/osi-pipeline-role"
area: "us-east-1"
serverless: false
index: "observability-otel-logs%{yyyy-MM-dd}"
# To get the values for the placeholders:
# 1. {OpenSearch_domain_endpoint}: You will discover the area endpoint by navigating to the Amazon Managed Opensearch managed clusters within the AWS Administration Console, after which clicking on the area.
# After acquiring the mandatory values, change the placeholders within the configuration with the precise values. Create a pipeline for traces
Full the next steps to create a pipeline for traces:
- Open the OpenSearch Service console.
- Select Ingestion pipelines.
- Select Create pipeline.
- Outline the pipeline configuration by getting into the next:
model: "2"
entry-pipeline:
supply:
otel_trace_source:
path: "/v1/traces"
processor:
- trace_peer_forwarder:
sink:
- pipeline:
title: "span-pipeline"
- pipeline:
title: "service-map-pipeline"
span-pipeline:
supply:
pipeline:
title: "entry-pipeline"
processor:
- otel_traces:
sink:
- opensearch:
index_type: "trace-analytics-raw"
hosts: ["{OpenSearch_domain_endpoint}"]
aws:
sts_role_arn: "arn:aws:iam::{accountId}:function/osi-pipeline-role"
area: "us-east-1"
service-map-pipeline:
supply:
pipeline:
title: "entry-pipeline"
processor:
- service_map:
sink:
- opensearch:
index_type: "trace-analytics-service-map"
hosts: ["{OpenSearch_domain_endpoint}"]
aws:
sts_role_arn: "arn:aws:iam::{accountId}:function/osi-pipeline-role"
area: "us-east-1"
# To get the values for the placeholders:
# 1. {OpenSearch_domain_endpoint}: You will discover the area endpoint by navigating to the Amazon Managed Opensearch managed clusters within the AWS Administration Console, after which clicking on the area. # 2. {accountId}: That is your AWS account ID. You will discover your account ID by clicking in your username within the top-right nook of the AWS Administration Console and choosing "My Account" from the dropdown menu.
# After acquiring the mandatory values, change the placeholders within the configuration with the precise values. Set up the OpenTelemetry demo utility in Amazon EKS
Use the EKS cluster you arrange earlier together with AWS CloudShell or one other device to finish these steps:
- Open the AWS Administration Console.
- Select the CloudShell icon within the prime navigation bar, or go on to the CloudShell console.
- Await the shell atmosphere to initialize—it comes preinstalled with widespread AWS Command Line Interface (AWS CLI) instruments.
Now you may full the next steps to put in the appliance.
- Clone the OpenTelemetry Demo repository:
git clone https://github.com/aws-samples/sample-correlation-opensearch-repository- Navigate to the Kubernetes listing:
cd deployment_files- Deploy the demo utility utilizing kubectl apply:
kubectl apply -f .- Use a load balancer to show the frontend service so you may attain the supply utility net URL:
kubectl expose deployment opentelemetry-demo-frontendproxy --type=LoadBalancer --name=frontendproxy- After you’ve gotten deployed the appliance, entry the frontend utility utilizing the load balancer on port 8080. Use your browser to go to http://<LoadBalancerIP>:8080/ to open the supply utility for OpenTelemetry.
By following these steps, you may efficiently set up and entry demo purposes in your EKS cluster.
Configure the OpenTelemetry Collector exporter for logs, traces, and metrics
The OpenTelemetry Collector is a device that manages the receiving, processing, and exporting of telemetry knowledge out of your utility to a goal repository.
On this step, we ship logs and traces to OpenSearch Service and metrics to Amazon Managed Prometheus. The OpenTelemetry Collector additionally works with common knowledge repositories like Jaeger and quite a lot of different open supply and business platforms. On this part, we embrace steps to configure the OpenTelemetry Collector in an EKS atmosphere. Then we deploy the demo utility and discover the OpenTelemetry exporters utilizing AWS Managed Options as a substitute of the open supply variations.
Full the next steps:
- Open the otel-collector-config ConfigMap in your most well-liked editor:
kubectl edit configmap opentelemetry-demo-otelcol -n otel-demo- Replace the exporters part with the next configuration (present the suitable Amazon Managed Service for Prometheus endpoint and OpenSearch Service log ingestion URLs):
exporters:
logging: {}
otlphttp/logs:
logs_endpoint: "<AWS_OPENSEARCH_LOG_INGESTION_URL>/v1/logs"
auth:
authenticator: sigv4auth
compression: none
otlphttp/traces:
traces_endpoint: "<AWS_OPENSEARCH_TRACE_INGESTION_URL>/v1/traces"
auth:
authenticator: sigv4auth
compression: none
prometheusremotewrite:
endpoint: "<AWS_MANAGED_PROMETHEUS_ENDPOINT>"
auth:
authenticator: sigv4auth - Find the extensions part and replace the IAM function ARN within the sigv4auth configuration:
sigv4auth:
assume_role:
arn: "arn:aws:iam::{accountId}:function/osi-pipeline-role"
sts_region: "us-east-1"
area: "us-east-1"
service: "osis"
# {accountId}: change accountID together with your account id- After updating the ConfigMap, restart the OpenTelemetry Collector deployment:
kubectl rollout restart deployment opentelemetry-demo-otelcol -n otel-demoWith these adjustments, the OpenTelemetry Collector will ship hint knowledge to the OpenSearch Service area, metrics knowledge to the AWS Managed Service for Prometheus endpoint, and log knowledge to the OpenSearch Service area.
Configure Amazon Managed Grafana
Earlier than you may visualize your logs and traces, you must configure OpenSearch Service as an information supply in your Amazon Managed Grafana workspace. This configuration is completed by way of the Amazon Managed Grafana console.
Configure the OpenSearch Service knowledge supply
Full the next steps to configure the OpenSearch Service knowledge supply:
- Open the Amazon Managed Grafana console.
- Choose your workspace and select the workspace URL to entry your Grafana occasion.
- Log in to your Amazon Managed Grafana occasion.
- From the facet menu, select the configuration (gear) icon.
- On the Configuration menu, select Knowledge Sources.
- Select Add knowledge supply.
- On the Add knowledge supply web page, choose OpenSearch Service from the record of obtainable knowledge sources.
- Within the Title area, enter a descriptive title for the info supply.
- Within the URL area, enter the URL (OpenSearch Service area endpoint) of your OpenSearch Service area, together with the protocol and port quantity.
- In case your OpenSearch cluster is configured with authentication, present the required credentials within the Consumer and Password
- If you wish to use a particular index sample for the info supply, you may specify it within the Index title area (For instance,
logstash-*). - Modify another settings as wanted, such because the Time area title and Time interval.
- Select Save & Check to confirm the connection to your OpenSearch cluster.
If the check is profitable, you need to see a inexperienced notification with the message “Knowledge supply is working.”
- Select Save to avoid wasting the info supply configuration.
- Repeat the identical steps for the OpenSearch logs and traces domains.
Configure the Prometheus knowledge supply
Full the next steps to configure the Prometheus knowledge supply:
- Open the Amazon Managed Grafana console.
- Choose your workspace and select the workspace URL to entry your Grafana occasion.
- Log in to your Amazon Managed Grafana occasion.
- From the facet menu, select the configuration (gear) icon.
- On the Configuration menu, select Knowledge Sources.
- Select Add knowledge supply.
- On the Add knowledge supply web page, choose Amazon Managed Prometheus from the record of obtainable knowledge sources.
- Within the Title area, enter a descriptive title for the info supply.
- The AWS Auth Supplier and Default Area fields needs to be mechanically populated primarily based in your Amazon Managed Grafana workspace configuration.
- Within the Workspace area, enter the ID or alias of your Amazon Managed Prometheus workspace.
- Select Save & Check to confirm the connection to your Amazon Managed Prometheus workspace.
If the check is profitable, you need to see a inexperienced notification with the message “Knowledge supply is working.”
- Select Save to avoid wasting the info supply configuration.
Create correlations in Amazon Managed Grafana
To determine connections between your logs and traces knowledge, you must arrange knowledge correlations in Amazon Managed Grafana. This lets you navigate seamlessly between associated logs and traces. Comply with these steps in your Amazon Managed Grafana workspace:
- Open the Amazon Managed Grafana console.
- Choose your workspace and select the workspace URL to entry your Grafana occasion.
- Within the Amazon Managed Grafana portal, on the Administration menu, select Plugins and Knowledge, and select Correlation.
- On the Arrange the goal for the correlation web page, underneath Goal, select your traces knowledge supply (OpenSearch Service, for instance,
otel-traces) from the dropdown record and outline the question that can execute when the hyperlink is adopted. You should use variables to question particular area values. For instance,traceId: ${__value.uncooked}.
- On the Arrange the goal for the correlation web page, select the log knowledge supply from the dropdown record, and enter the sphere title to be linked or correlated with the traces knowledge supply within the OpenSearch Service knowledge supply. For instance,
traceID.
- Select Save to finish the correlation configuration.
- Repeat the steps to create a correlation between metrics on Prometheus to logs in OpenSearch Service.
Validate outcomes
In Amazon Managed Grafana, utilizing the Prometheus knowledge supply, find the specified occasion for correlation. The occasion ID will likely be displayed as a hyperlink. Comply with the hyperlink to open the corresponding log particulars in a panel on the best facet of the web page.
With the logs to traces correlation configured, you may entry hint info instantly from the logs web page. Select traces on the log particulars panel to view the corresponding hint knowledge.
The next screenshot demonstrates the node graph visualization displaying the correlation stream: occasion metrics to logs to traces.
Clear up
Take away the infrastructure for this answer when not in use to keep away from incurring pointless prices.
Conclusion
On this submit, we confirmed the way to use correlation as a useful device to achieve perception into observability knowledge saved in varied shops.
Separating logs and traces into devoted domains supplies the next advantages:
- Higher useful resource allocation and scaling primarily based on completely different workload patterns
- Unbiased efficiency optimization for every knowledge kind
- Simplified value monitoring and administration
- Enhanced safety management with separate entry insurance policies
You should use this answer as a reference to construct a scalable observability answer on your enterprise to detect, examine, and remediate issues sooner. This capacity, when used alongside next-generation synthetic intelligence and machine studying (AI/ML), helps to not solely proactively react however predict and forestall issues earlier than they happen. You’ll be able to be taught extra about AI/ML with AWS.
In regards to the Authors
Balaji Mohan is a Senior Supply Marketing consultant specializing in utility and knowledge modernization to the cloud. His business-first strategy supplies seamless transitions, aligning expertise with organizational targets. Utilizing cloud-centered architectures, he delivers scalable, agile, and cost-effective options, driving innovation and development.
Senthil Ramasamy is a Senior Database Marketing consultant at Amazon Net Providers. He works with AWS clients to offer steerage and technical help on database providers, serving to them with database migrations to the AWS Cloud and bettering the worth of their options when utilizing AWS.
Muthu Pitchaimani is a Search Specialist with Amazon OpenSearch Service. He builds large-scale search purposes and options. Muthu is within the subjects of networking and safety, and relies out of Austin, Texas.
