See your system without running another cluster.
DeepCube is a managed service, so there is no cluster to size, shard or tune. It runs next to Elastic on your existing OpenTelemetry.
What If Observability Just Worked?
Point your OpenTelemetry Collector at our OTLP endpoint and your system appears as a live 3D map you move through. That's spatial observability.
Is DeepCube an Elastic alternative?
For seeing traces in a live map with no cluster to run, yes. For deep log search, alerting, paging and infrastructure monitoring, no. Keep Elastic for those and run DeepCube next to it. Visualize takes 200 traces and logs per second for the whole grid, so check that against your volume. DeepCube is a Windows app today, with an alpha on macOS. It shows a quiet consumer or an unanswered call as a gap, but it does not yet flag a service that stops reporting.
Intro
Elastic Observability is built on Elasticsearch. On self-managed and Elastic Cloud Hosted, running it comes with operational weight: nodes, shards and capacity stay with you. Elastic's pricing page asks "Want Elastic to manage nodes and shards for you?" and points to its other offerings (source, verified 2026-11-06). On Elastic Cloud Serverless, by contrast, Elastic says it "automatically manages" cluster scaling, node management and "Shard distribution and replication" (source, verified 2026-11-14).
DeepCube is the product; spatial observability is the category. We are not an APM company, we are a spatial observability company.
Observe your system. Not your observability platform.
The Infrastructure Project Problem
Elastic Observability runs on Elasticsearch, and on these deployments you operate it.
- Cluster management, on self-managed and Cloud Hosted: nodes, shards and capacity (source, verified 2026-11-06).
- Operational overhead that stays with you on those deployments. Elastic Cloud Serverless is the option where Elastic carries it: "no cluster management, no need to manage underlying infrastructure, no upgrades" (source, verified 2026-11-14).
- DeepCube is managed: there is no cluster to run for the map.
No clusters. No shards. Just your system.
What If Observability Just Worked?
No clusters. No shards. No ILM policies. Just your system.
- Managed platform: Point your OpenTelemetry Collector at our OTLP endpoint and your system appears as a spatial topology.
- Services, dependencies, and telemetry in a single 3D environment. No infrastructure to manage.
- Your engineers focus on their systems. We handle the observability platform.
- Phantoms: when a call goes out and nothing ever reports back from the other side, or a consumer goes quiet while messages keep arriving, DeepCube draws the missing side where it should be, so what is missing shows up in the space. It does not yet flag a service that stops reporting.
Zero clusters. Zero shards. Your system in 3D.
Architecture: How We Differ
OpenTelemetry only vs Elastic Agent or EDOT.
DeepCube is built exclusively for observability. No cluster management, no shard allocation, no ILM policies to configure. Your engineers spend time understanding their systems, not operating the platform that monitors them.
Elastic has 550+ integrations (source, verified 2026-11-06), search and query capabilities for log analysis, and a flexible data platform that spans observability, security, and search. ES|QL and Elasticsearch's full-text, semantic, and vector search support ad hoc log investigation (languages, search approaches, verified 2026-11-10). For teams deeply invested in the Elastic ecosystem, the breadth of the platform is a genuine advantage.
Elastic is also candid about what its OpenTelemetry-native path gives up against its own classic stack, and it is worth reading before you count integrations. Their compatibility page states that "Many prebuilt Elastic integrations and dashboards are designed for ECS-formatted data and may not work as expected with the OpenTelemetry semantic conventions without customization", that its OTel-native browser instrumentation, EDOT Browser, is "in technical preview", so "continue using the classic Elastic APM browser agent for production RUM", that Universal profiling "is currently only supported in the classic stack", and that "Elastic OpenTelemetry does not provide managed TBS" (source, verified 2026-11-06).
| Aspect | DeepCube | Elastic Observability |
|---|---|---|
| Instrumentation | Standard OpenTelemetry SDKs only | EDOT SDKs (Elastic's OpenTelemetry distributions), or classic Elastic APM agents (source, verified 2026-11-09) |
| Data Ingestion | OTLP over gRPC | EDOT or Elastic Agent, the Managed OTLP Endpoint, or APM Server (legacy) (source, verified 2026-11-10) |
| Visualization | 3D spatial topology | Kibana dashboards, charts, maps, plus an APM Service Map, which Elastic 9.5 carries into dashboards ("Adds the APM Service Map as a dashboard embeddable panel") and alerts ("Adds an APM Service Map preview to the Observability Alerts page for service topology context") (release notes, verified 2026-11-06). Elastic's own map controls offer a choice between "Horizontal and Vertical" layout (source, verified 2026-11-06) |
| Operational Overhead | Fully managed SaaS | Self-managed and Cloud Hosted: capacity, shard counts and ILM policies are yours to set. Serverless removes these, but its entry Logs Essentials tier leaves out "Machine learning", "Unified agentic AI and scripted workflow automation" and the "Fully curated experiences for analysis of logs, metrics, traces, and synthetic tests"; those are Observability Complete only (pricing, verified 2026-11-06) |
| Integrations | OTel ecosystem (community-driven) | 550+ out-of-the-box integrations (source, verified 2026-11-06). Elastic qualifies that advantage itself for an OTel-first buyer: prebuilt integrations and dashboards "may not work as expected with the OpenTelemetry semantic conventions without customization" (OTel limitations, verified 2026-11-06) |
| Search / Query | Structured queries + spatial navigation | ES|QL, KQL, Query DSL; full-text, vector, semantic and hybrid search (languages, search approaches, verified 2026-11-10) |
In DeepCube, Tessa starts from the spatial view and works through your traces, logs and code on her own, then brings you the answer. You own the outcome.
AI: Tessa vs Elastic Agent Builder
In DeepCube, Tessa connects the telemetry of the environment you are viewing to your connected codebase. In DeepCube, add your repositories as workspaces, and Tessa can help you follow a production issue from the 3D topology to the source file. In DeepCube, Tessa accesses your connected codebase workspace, diagnoses from 3D topology, proposes the fix. You own the outcome.
Elastic scopes its AI Assistant to your Elastic data: "The AI Assistant is an integration with a large language model (LLM) that helps you understand, analyze, and interact with your Elastic data", and the visualization it offers is "Create time-series charts and distribution graphs from your Elasticsearch data" (source, verified 2026-11-06). It is also now deprecated, with the Elastic AI Agent as its stated replacement. Valuable capabilities, but on self-managed and Elastic Cloud Hosted subscriptions it is gated to the Enterprise tier ("As low as" $184/mo; entry pricing "As low as" $99/mo Standard) (tier gating: elastic.co/subscriptions, which marks Elastic AI Assistant Enterprise-only across Elasticsearch, Observability and Security; price: elastic.co/pricing/cloud-hosted; both verified 2026-11-06). Elastic Cloud Serverless tiers differently: there the Elastic Managed LLM is scoped to Observability Complete projects and metered per token. Elastic reaches your code through a GitHub connector, a coding agent you bring and, in private preview, Elastic nightshift AI SRE (see Code Fixes). AI that works in your code is common now, and, in DeepCube, Tessa does it too; the difference is where Tessa starts, the spatial view of your running system, and that it is in the per-node price.
| Capability | Tessa (DeepCube) | Elastic Agent Builder (GA 2026-01-22) |
|---|---|---|
| Codebase Access | In DeepCube, your local checkout: read, search, propose edits | Through connectors, not built-in tools. Elastic's GitHub connector searches code, reads file contents and can open pull requests (source, verified 2026-11-06). The tools themselves stay on Elastic data: "Tools can be thought of as functions: modular, reusable actions that agents can call to interact with your Elasticsearch data", and "Tools enable agents to work with Elasticsearch data" (tools, verified 2026-11-06). Their MCP server runs outward by their own description: it "provides a standardized interface for external MCP hosts to access Elastic Agent Builder tools. For example, you can run an ES|QL query against your data from Claude Desktop without opening Kibana" (source, verified 2026-11-06) |
| Spatial Context | In DeepCube: 3D topology + traces | Kibana dashboards and ES|QL queries, plus a real service topology: Agent Builder ships observability.get_service_topology, which "Retrieves the service topology (dependency graph) for a service, including RED metrics (latency, throughput, and error rate) per connection", inside a family Elastic scopes as "Observability tools provide specialized capabilities for monitoring applications, infrastructure, and logs" (built-in tools reference, verified 2026-11-06). The rendered Service Map is, by Elastic's own caveat, incomplete: "if a service isn't instrumented, or a traceparent header isn't being propagated to it, distributed tracing will not work, and the connection will not be drawn on the map" (service map, verified 2026-11-06). A dependency graph, not a space you can move through |
| Code Fixes | In DeepCube, Tessa proposes the fix in your workspace. You own the outcome | Elastic nightshift AI SRE, in private preview, recommends a fix as a workflow or connector action, or as commands for a terminal or coding agent (source, verified 2026-11-06). Elastic's agent-skills program ships "official, open-source skill packages that help AI coding agents work with the Elastic stack", which "give agents like Claude Code, Cursor, GitHub Copilot, and others the specialized knowledge they need to perform Elastic-specific tasks more accurately and efficiently" (source, verified 2026-11-06). Your coding agent learns Elastic through these skills |
| Availability | Included in the price of every plan. No AI seat required, no per-query AI meter. Optional Tessa Personal seats (Tessa E5, E25 and E50 on the pricing page) add capacity. Model quality scales with the plan. | Enterprise tier only on Cloud Hosted, and Elastic states that tier as a floor rather than a price: "As low as $184 per month" (tier matrix, pricing, verified 2026-11-06) |
| Focus | Diagnosis + code fix with spatial context, in DeepCube | Agent Builder: built-in agents and tools for observability, plus custom ones (source, verified 2026-11-10) |
| Accountability Model | In DeepCube, Tessa works through your traces, logs and code on her own, then brings you the answer. You own the outcome | Agent Builder uses human-in-the-loop prompts for actions inside Elastic, and Elastic 9.5 added Workflows steps to "approve, reject, or submit structured input" in Kibana or through signed links and Slack notifications (Kibana release notes, verified 2026-11-06). The prompts are scoped: "Human-in-the-loop prompts work only in interactive Elastic Agent Builder conversations", and "A sub-agent execution can't collect a response to these prompts" (source, verified 2026-11-06) |
Note: Elastic Agent Builder, generally available since 2026-01-22 ("Agent Builder is now GA", Elastic Search Labs, verified 2026-11-06), is now the strategic direction; the original Elastic AI Assistant is now deprecated, and Elastic's own docs banner reads "The Elastic AI Agent is now the default chat experience"; both are Enterprise-gated. Agent Builder and Workflows are metered on Serverless: 10,000 executions free, then as low as $0.025 and $0.0108 per execution respectively, and both are available only on Observability Complete projects (source, pricing, verified 2026-11-06).
One per-node price. AI in every plan.
What you would pay, and for what
Elastic's pricing model combines resource-based cloud pricing with tier-gated features. This table shows how each one bills and what each price covers:
- AI is in every tier: Tessa is in the price of every DeepCube plan, including the free Start tier; higher plans add capacity, model quality and features. On Elastic Cloud Hosted, Elastic gates anomaly detection and supervised-learning ML to Platinum (as low as $131/mo) and the AI Assistant to Enterprise (as low as $184/mo) (tier matrix, pricing, verified 2026-11-06).
- One primary meter: DeepCube charges per node, within the plan limits listed below. Elastic bills across GB ingest and GB retention on Serverless (or instance RAM-hours on Cloud Hosted), egress, LLM tokens, synthetic test runs, cross-project search, and per-execution Agent Builder ($0.025) and Workflows ($0.0108) charges (effective 2026-05-01) (serverless pricing, hosted pricing, verified 2026-11-06).
- Support by tier: Support varies by plan on both products. On Elastic Cloud Hosted, support is tier-gated: the Standard tier (as low as $99/mo) is web-only with a 3-business-day target response covering the Cloud platform only, and 24/7/365 support starts at Platinum (as low as $131/mo) (pricing, verified 2026-11-06).
- Their topology view is not in the entry tier either: Elastic's Cloud Hosted subscription matrix marks Service maps unavailable on Standard and Gold, and available from Platinum up (Cloud tier matrix, verified 2026-11-09).
- The AI can start billable ML nodes: "if you have machine learning autoscaling enabled, machine learning nodes automatically start when using the knowledge base and AI Assistant", and Elastic says that incurs additional costs (source, verified 2026-11-06).
- The tier prices are floors, not prices: Elastic lists Standard, Platinum and Enterprise as "As low as" $99, $131 and $184 per month; each is based on a 120 GB, 2-zone config, and larger deployments cost more (pricing, verified 2026-11-06).
- Predictable math: Node count sets the bill, with a volume discount on Visualize.
| Line Item | DeepCube | Elastic Observability |
|---|---|---|
| Bills on | Per application instance (a node) | Cloud Hosted: instance usage, in GB of RAM per hour. Serverless: GB ingested and GB retained per month (billing, serverless pricing, verified 2026-11-06) |
| List rate | Visualize: $20 per node, with a volume discount | Cloud Hosted: "As low as" $99/mo (Standard tier) entry; "As low as" $184/mo (Enterprise tier, required for the AI Assistant), each based on a 120 GB, 2-zone config (pricing, tier matrix, verified 2026-11-06) |
| Ingest and retention (Serverless) | In the node price, within plan limits | Logs and traces: ingest $0.07-0.09/GB, retention $0.017-0.019/GB/mo. Metrics have their own rates since 2026-07-01: on Complete, $0.023/GB ingest and $0.005/GB/mo retention (source, verified 2026-11-06) |
| AI Assistant | Included (Tessa) | Enterprise tier required on Cloud Hosted ("As low as" $184/mo), plus ML nodes if ML autoscaling is on (tier matrix, verified 2026-11-06) |
| ML (Anomaly Detection) | Not a DeepCube feature | Platinum+ required on Cloud Hosted ("As low as" $131/mo); anomaly detection and supervised learning (tier matrix, verified 2026-11-06) |
| Service Maps | Included (3D spatial topology) | Platinum+ required on Cloud Hosted; not available on Standard or Gold (Cloud tier matrix, verified 2026-11-09) |
| In the price | Logs and Tessa | Cloud Hosted: features and support by tier. Enterprise lists Agent Builder and the AI Assistant; 24/7/365 support starts at Platinum. Serverless: 50 GB of egress free, then $0.05/GB (pricing, serverless pricing, verified 2026-11-06) |
| Not in the price | Infrastructure monitoring, alerting and paging. Optional: Tessa Personal seats (Tessa E5, E25 and E50), support hours. A $99.99 fee applies when you change plan. | Serverless: "Synthetic Monitoring, GenAI capabilities, and cross-project querying are available as add-ons." (source, verified 2026-11-06) |
| Plan limits | Visualize: 200 traces and logs per second for the whole grid, 30-day retention, shared infrastructure. Start: 50 per second, 7-day retention. Telemetry above the limit can be dropped, and your collector will not show an error. DeepCube is a Windows app today, with an alpha on macOS. | See its pricing page |
| What it is for | Runs next to your current tool for tracing failures. No self-serve plan is meant to stand in for it today. Analyze and Fuse are available now through sales, with node minimums of 25 and 50. | Cloud Hosted is "a public cloud managed service", "delivering search, AI-driven security, and cloud monitoring capabilities" (source, verified 2026-11-06) |
| Your number | Count your application instances, then see deepcube.ai/pricing | elastic.co/pricing |
Elastic pricing based on published rates from elastic.co/pricing, verified 2026-11-06. Actual costs vary by resource consumption, tier, and deployment model. DeepCube pricing from deepcube.ai/pricing. All prices USD.
It runs next to what you have
Run DeepCube next to your current tool.
Traces and metrics to two destinations, no re-instrumentation if you already run OTel.
- Dual-destination: Your OTel Collector fans out traces and metrics to both DeepCube and Elastic simultaneously.
- No code changes: If you already use OTel SDKs, your instrumentation does not change.
- If you use Elastic Agent: Keep Elastic Agent where it is. DeepCube needs OpenTelemetry, so add OTel instrumentation to the services you want in the map.
- Keep Elastic for logs: Use DeepCube for traces and spatial visualization while keeping Elastic for log search.
- Easy exit: If DeepCube is not right for you, change one endpoint and its API key header. Your instrumentation stays exactly the same.
Dual-Destination Collector Config
receivers:
otlp:
protocols:
grpc:
http:
exporters:
otlp/deepcube:
endpoint: "https://otlp.deepcube.ai"
headers:
API-Key: "YOUR-API-KEY"
otlp/elastic:
endpoint: "${ELASTIC_OTLP_ENDPOINT}"
headers:
Authorization: "ApiKey ${ELASTIC_API_KEY}"
service:
pipelines:
traces:
receivers: [otlp]
exporters: [otlp/deepcube, otlp/elastic]
metrics:
receivers: [otlp]
exporters: [otlp/deepcube, otlp/elastic]
logs:
receivers: [otlp]
exporters: [otlp/deepcube]
Standard OTel Collector config. Traces and metrics go to both destinations. Logs go to DeepCube only, so leave your existing log route in place. Endpoint and ApiKey header follow Elastic's Managed OTLP Endpoint docs (source, verified 2026-11-10). No code changes required if you already run OTel.
Ready to See the Difference?
Start free with DeepCube. Your OTel instrumentation stays as it is: send it through a Collector or any OTLP/gRPC exporter.