> For the complete documentation index, see [llms.txt](https://developer.harness.io/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://developer.harness.io/continuous-delivery/use-continuous-delivery/verify-deployments/verify-deployments-with-the-verify-step.md).
# Verification Overview
Get an overview of Harness Continuous Verification with AI Verify
Continuous Verification (CV) is a step in your deployment pipeline that validates a deployment automatically. It integrates with the Application Performance Monitoring (APM) and log management tools you already use, analyzes metrics and logs collected during and after a deployment, and triggers a rollback if it detects anomalies.
## Why use Continuous Verification
Without automated verification, teams typically rely on someone watching dashboards after a deployment, or waiting for users to report problems, to catch regressions. Continuous Verification moves that check into the pipeline itself:
* **Faster detection**: CV analyzes metrics and logs as soon as they're available after a deployment, instead of waiting for a person to notice a problem or a customer to report one.
* **Consistent criteria**: Every deployment is evaluated the same way, either against a historical baseline and thresholds you configure, or, with AI Verify (v2), against statistical anomaly detection, so the outcome doesn't depend on who's watching at the time.
* **Automatic rollback**: When CV detects anomalies, it can fail the step and trigger a rollback, limiting the blast radius of a bad deployment.
* **Fits your existing observability stack**: CV integrates with the APM and log management tools your teams already use for monitoring, so you don't need to adopt a separate platform just for verification.
## Choose a Verify step
Harness offers two Verify steps for Continuous Verification:
{% hint style="info" %}
**AI Verify (v1)**
The original Verify step. Requires a Monitored Service, manually configured health sources, and threshold tuning. Applies machine learning to compare each deployment against a historical baseline.
Go to [AI Verify (v1) overview](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv/ai-verify-v1/overview.md) for deployment strategies, sensitivity, duration, fail-fast thresholds, health sources, and the Service Instance Identifier (SII).
{% endhint %}
{% hint style="info" %}
**AI Verify (v2)**
A newer, agentic Verify step. An LLM-powered Configuration Agent discovers and configures health sources for you, and statistical anomaly detection replaces manual threshold tuning.
Go to [AI Verify (v2) overview](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv/configure-verify-step-with-ai/overview.md) for components, how it differs from AI Verify (v1), and how to get started.
{% endhint %}
## What's supported
Each Verify step supports a different set of Application Performance Monitoring (APM) and log management tools.
### AI Verify (v1)
A health source monitors health trends in a service by analyzing metrics from APM tools and logs from log management tools. AI Verify (v1) supports the following APM and log management tools:
#### APM tools
Harness CV supports the following APM tools:
* AppDynamics
* Amazon CloudWatch
* Custom health source
* Datadog
* Dynatrace
* Google Cloud Operations (formerly Stackdriver)
* New Relic
* Prometheus (includes support for AWS Prometheus)
* Splunk Observability (formerly SignalFx)
* Sumo Logic
Harness supports most of the popular APM tools, but there may be instances where Harness don't have a native connector. Using the **Harness Custom Health Source** feature, you can integrate such APM tools with Harness.
#### Log management tools
Harness CV supports the following log management tools:
* Datadog
* Elasticsearch
* Google Cloud Operations (formerly Stackdriver)
* Grafana Loki
* Splunk
* Sumo Logic
Harness supports most of the popular log management tools, but there may be instances where Harness don't have a native connector. Using the **Harness Custom Health Source** feature, you can integrate such log management tools with Harness.
### AI Verify (v2)
AI Verify (v2) uses AI-assisted health source configuration, where a Configuration Agent discovers and configures health sources for you instead of manual setup. Platform support for AI Verify (v2) is currently narrower than for AI Verify (v1):
* **Metrics**: Datadog, with AI-assisted configuration. New Relic Metrics is also supported with manual setup.
* **Logs**: Dynatrace, with AI-assisted configuration. Splunk Logs is also supported with manual setup.
* **All other observability platforms** (Prometheus, AppDynamics, and others): not yet supported for AI Verify (v2). Use AI Verify (v1) for these platforms.
Go to [AI Verify (v2) overview](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv/configure-verify-step-with-ai/overview.md#limitations-and-considerations) for more on platform support and limitations.
{% hint style="info" %}
**NOTE**
When configuring an APM tool or a log management tool in Harness CV and constructing a query, consider the following:
* The query result should be below 100 logs per minute. If the logs exceed this limit, Harness performs random sampling for processing.
* The query response should be a single time series.
{% endhint %}
{% hint style="warning" %}
When using Datadog as a health source in Harness CV, be aware that the default DataDog rate limit is 360 API calls per hour, which translates to 6 API calls per minute. Since Continuous Verification (CV) steps in Harness make 1 API call per pod, this effectively restricts the maximum number of pods or instances to 6.
For example, if you test a deployment with just 1 instance (replica), the CV step will work as expected. However, if you run the same workflow in a production or staging environment where the replica count exceeds 20, the CV step is likely to fail due to the API rate limit.
If needed, you can request an increase in the rate limit from DataDog to accommodate larger deployments.
{% endhint %}
To learn how to add a health source, go to [Configure verification](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv.md).
For information about what's supported for other Harness modules and the Harness Platform overall, go to [Supported platforms and technologies](/harness-platform/new-to-harness-platform/platform-whats-supported.md).
## FAQs
Go to [Continuous Verification FAQs](/continuous-delivery/use-continuous-delivery/verify-deployments/continuous-verification-faqs.md) for frequently asked questions about Continuous Verification.
## Next steps
* [AI Verify (v1) overview](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv/ai-verify-v1/overview.md): Concepts, configuration, and health sources for the original Verify step.
* [AI Verify (v2) overview](/continuous-delivery/use-continuous-delivery/verify-deployments/configure-cv/configure-verify-step-with-ai/overview.md): Concepts and configuration for the agentic Verify step.
* [Understand Verification Results](/continuous-delivery/use-continuous-delivery/verify-deployments/cv-results.md): Interpret verification output.
{% @harness-feedback/feedback %}