Discovery

Reply CMP Discovery maps your cloud estate across providers and keeps a live CMDB of resources, relationships, and changes over time.

Note

Discovery is read‑only. It inventories resources via provider APIs and never modifies your cloud configuration. To create resources, use Provisioning.

Quick start

  1. Open Discovery from the left navigation.

  2. Use the category tabs (Compute, Serverless, Database, Network, Storage, Other) or stay on All.

  3. Filter by Provider, Connection, Tags, Groups, or Type.

  4. Click a resource to open its details (overview, links, properties, graph, history).

Tip

Use the search bar to find resources by name, tag value, or type. Combine with filters to narrow large estates quickly.

How it works

  • Multi‑cloud: Unifies Azure, AWS, and GCP resources in a single view.

  • CMDB: Stores configuration, relationships, and a time‑stamped change history for each resource.

  • Relationships: Builds a graph of dependencies so you can trace upstream/downstream impact.

  • Data collection: Uses your configured connections and provider monitoring/management APIs.

Note

Manual refresh is available per connection. Scheduled crawls keep the CMDB current; you can trigger a refresh whenever you need the latest state.

Resource details

Open any resource to see:

Overview

  • Name, Region, Type, First discovered / Last updated

  • Connection card

  • Provider link (quick jump to the provider console)

  • Provisioned link:

    • If the resource was created by Reply CMP, you’ll see the deployment/resource mapping with a link back to Provisioning.

    • If not, it’s labeled “Not provisioned via CMP.”

Properties explorer

  • Full configuration as JSON with inline search.

  • Copy snippets for reviews or change requests.

Resource graph

  • Visualizes relationships (for example, NIC → VM → Subnet → VNet).

  • Navigate the graph to inspect neighbors without leaving the page.

History

  • A timeline of changes with the option to compare two snapshots side‑by‑side.

  • Highlighted diffs help you spot what changed and when.

  • AI‑generated rollback suggestions can provide example commands (for example, az CLI or Terraform) you can run in your environment after review.

Warning

Rollback suggestions do not change your cloud automatically. Review carefully before executing in your environment.

Provider-Specific Discovery Methods

Azure

Method: Azure Resource Graph (ARG)

Azure Resource Graph provides a fast, efficient way to query resources at scale across subscriptions. Reply CMP uses ARG to retrieve all resources and resource groups within your configured subscription.

What’s retrieved:

  • All Azure resources (VMs, disks, storage accounts, databases, networking components, etc.)

  • Resource properties: ID, name, type, region, tags, resource group

  • Resource metadata and configuration details

Typical refresh time: 1-3 minutes (ARG is highly optimized for large-scale queries)

Advantages:

  • Near-instant queries across thousands of resources

  • Subscription-scoped visibility

  • Consistent API performance regardless of estate size

AWS

Method: AWS Resource Explorer 2 + AWS Cloud Control API

AWS Resource Explorer 2 provides a unified index of resources, which Reply CMP then enriches with detailed metadata using the Cloud Control API.

What’s retrieved:

  1. Resource Explorer 2: List of all resources by ARN, region, type, and tags

  2. Cloud Control API: Detailed metadata enrichment

    • Prefers GetResource (more detail) over ListResources (faster but less detail)

    • Retrieves runtime state, configuration, and resource-specific properties

Prerequisites:

  • Resource Explorer 2 must be enabled in your account and configured with a global view

  • The connection’s IAM user must have permissions to query Resource Explorer and Cloud Control

Typical refresh time: 5-10 minutes (depends on resource count and API rate limits)

Resource grouping: Resources are organized by type and region for efficient browsing

Note

Resource Explorer 2 must be enabled in the connection’s region. If you don’t see resources, verify it’s active in AWS console.

GCP

Method: Cloud Asset Inventory + Service-Specific APIs

Google Cloud’s Asset Inventory provides a centralized, strongly consistent view of all project assets, which Reply CMP supplements with service-specific APIs for runtime details.

What’s retrieved:

  1. Cloud Asset Inventory: Unified list of all project assets

    • Compute instances, persistent disks, Cloud SQL databases

    • GKE clusters, Cloud Storage buckets, Cloud Run services

    • Pub/Sub topics, Cloud Functions, and more

  2. Service-Specific APIs: Additional metadata for:

    • Compute Engine (instance status, machine type, network config)

    • Cloud SQL Admin (database version, tier, backup config)

    • GKE (cluster status, node pools, version)

    • Cloud Storage (bucket location, storage class)

    • Pub/Sub (topic subscriptions, message retention)

    • Cloud Run (service revision, traffic split)

    • Cloud Functions (runtime, trigger type)

Prerequisites:

  • Cloud Asset Inventory API must be enabled in the project

  • Service account must have roles/cloudasset.viewer and appropriate service-level read permissions

Typical refresh time: 3-8 minutes (depends on asset count and API quotas)

Advantages:

  • Centralized, strongly consistent asset view

  • Comprehensive coverage across all GCP services

  • Efficient querying with filtering and relationship traversal

Data freshness

  • Discovery updates run on schedule (twice per day) and on demand (per connection).

  • Adds new resources, updates changed ones, and marks deleted resources accordingly.

  • Provider API limits and large estates can extend crawl duration; recently changed items may appear after the next refresh.

Troubleshooting

  • Resource missing: Ensure the correct connection and region are selected; run a manual refresh.

  • Empty properties: Some services require specific permissions; verify the connection’s role can read resource details.

  • Relationships incomplete: The graph improves after the first full crawl; refresh if dependencies were added recently.

  • “Not provisioned via CMP”: The item exists in your cloud but wasn’t deployed through Reply CMP (still fully discoverable).

FAQ

Does Discovery affect my resources?
No. Discovery is read‑only and uses provider APIs.

Can I export the inventory?
Yes. Use Export from the list view to download the filtered set.

How far back does history go?
From the moment the resource was first discovered by Reply CMP; each crawl adds a new snapshot when changes are detected.

Where are costs?
Costs and budgets live in FinOps. Use the resource name or tags there to analyze spend.

How often should I refresh?
Resources are refreshed automatically twice per day; trigger a manual refresh when making frequent changes or before audits.

Glossary

  • CMDB: Configuration Management Database storing resources, relationships, and change history.

  • Connection: The configured binding to a provider scope (subscription/account/project).

  • Relationship graph: The dependency map built from discovered links.

  • Snapshot: A version of resource properties captured at a specific time during discovery.