Dynatrace vs AppDynamics: Which APM Should You Standardize On?

The alert fired at 2:47 a.m. A payment service error rate spiking across three Kubernetes clusters. The on-call engineer opened the observability platform and waited. One tool traced the failure to a misconfigured database connection in eight seconds and pointed at the exact pod. The other needed twenty minutes of digging through business-transaction dashboards before anyone knew where to look. Both teams were running enterprise APM. Nothing about the experience was the same.

That gap is the real dynatrace vs appdynamics difference, and it begins long before the alert fires. At its core, this is a choice between a consumption-based, causal-AI full-stack platform and a business-transaction APM anchored inside the Cisco ecosystem. This guide examines the architectural, operational, and financial trade-offs that drive this decision at enterprise scale, using verified pricing data and documented deployment models rather than vendor marketing claims.

At-a-Glance: Dynatrace vs AppDynamics

Criterion	Dynatrace	AppDynamics (Cisco)
Instrumentation model	Single OneAgent per host; full-stack auto-instrumentation with no per-runtime agent swap-outs	Separate App Agents per runtime (Java, .NET, Node.js, PHP, etc.) plus separate Machine/Infrastructure Agent
AI approach	Davis causal AI: deterministic, topology-aware root-cause isolation with automated anomaly detection	Business iQ / Cognition Engine: business-transaction correlation maps application performance to business KPIs
Observability scope	APM, infrastructure, logs, digital experience (RUM/synthetic), security (RASP, vulnerability analytics, posture) on one platform	APM core plus Business iQ, Database Visibility, End-User Monitoring; broader stack coverage via Cisco Full-Stack Observability integrations (ThousandEyes, Intersight)
Pricing model	Consumption-based DPS (Dynatrace Platform Subscription); published public rate card; prepaid annual commit, no overcommit penalties	Quote-based agent/CPU licensing; no public rate card; sold as annual subscription, often bundled into Cisco FSO
Kubernetes / cloud-native maturity	Operator-based deployment; per-pod billing option; service-mesh and serverless coverage; Smartscape auto-topology	Kubernetes support available; historically stronger on VM/bare-metal estates; Cisco FSO adds broader cloud-native integrations
OpenTelemetry support	Ingests OTel traces, metrics, and logs natively; OTel data processed in Grail lakehouse alongside proprietary signals	Supports OTel for trace ingest; deeper instrumentation still relies on proprietary App Agents for full feature coverage
Deployment options	SaaS-first; Dynatrace Managed (self-hosted) available for regulated environments; air-gapped support is limited	SaaS, on-premises, and hybrid; historically strong on-prem footprint; better fit for air-gapped or heavily regulated deployments
Ecosystem and lock-in	Dynatrace platform; third-party integrations via 715+ supported technologies; OTel reduces proprietary dependency on signals layer	Cisco bundle play: AppDynamics, ThousandEyes, Intersight, Meraki; significant ecosystem leverage for existing Cisco shops, but deep Cisco dependency

Instrumentation and Deployment: OneAgent vs. App and Machine Agents

Instrumentation architecture determines day-one deployment effort, ongoing maintenance burden, and the ceiling of what the platform can auto-discover. The two tools take meaningfully different approaches.

Dynatrace OneAgent: Single Agent, Full-Stack Auto-Instrumentation

Dynatrace's OneAgent installs once per host and instruments everything running on that host: JVM processes, .NET CLR, Node.js, Go, PHP, databases, web servers, network interfaces, and OS metrics. There is no per-runtime agent to select or version-match. OneAgent uses kernel-level hooks and bytecode instrumentation to capture traces, metrics, and logs automatically, without code changes. In Kubernetes environments, the Dynatrace Operator deploys OneAgent as a DaemonSet or injects instrumentation at the pod level via the CSI driver, enabling the per-pod billing tier ($1.40/mo per pod as of mid-2026, per the Dynatrace public rate card).

The single-agent model simplifies fleet management at scale. Rolling out instrumentation to 500 hosts is one Ansible task or Helm chart, not a matrix of agent versions per language runtime. The trade-off is that OneAgent is a privileged process with broad host access, which adds a surface-area consideration in high-security environments and requires approval workflows in some regulated industries.

AppDynamics: Separate App Agents and Machine Agents

AppDynamics uses a tiered agent model. App Agents are runtime-specific: a Java App Agent, a .NET App Agent, a Node.js App Agent, and so on. Each is a separate artifact with its own release cadence, configuration, and compatibility matrix. The Machine Agent (also called the Infrastructure Agent) runs separately to collect host-level metrics and connects to the Controller. A Database Agent covers database monitoring, and an End-User Monitoring (EUM) agent handles browser and mobile instrumentation.

This architecture gives AppDynamics precise control over what gets instrumented and how deeply, which matters in environments where not every process should carry full tracing overhead. The operational cost is higher: agent selection, version pinning, and compatibility testing scale with the number of distinct runtimes in the estate. Teams running polyglot microservices across dozens of services commonly report that AppDynamics agent lifecycle management becomes a recurring operational task. Under Cisco ownership, agent deployment has been increasingly streamlined via Cisco's automation tooling, but the multi-agent model remains architecturally distinct from OneAgent.

AI and Root-Cause Analysis: Davis vs. Cognition Engine and Business iQ

Both platforms use the term "AI," but the underlying approaches target different problems. Understanding that difference is essential before evaluating either tool in a POC.

Dynatrace Davis: Deterministic Causal AI and Automated RCA

Davis, Dynatrace's AI engine, is deterministic and topology-aware. It operates against Dynatrace's Smartscape, a continuously updated real-time topology map of every entity (host, process, service, pod, database) and every dependency relationship between them. When an anomaly is detected, Davis does not produce a ranked list of probable causes. Instead, it traverses the dependency graph, identifies the causal entity (the one whose degradation propagates to the observed symptom), and surfaces a single root-cause hypothesis with a confidence score and a direct link to the affected entity.

This approach produces low-alert noise in environments where Smartscape is well-populated, typically within minutes of OneAgent deployment. The eight-second RCA in the scenario at the top of this article reflects a Davis-style answer: one affected pod, one probable cause, surfaced automatically. The architectural constraint is that Davis's accuracy degrades if Smartscape is incomplete, which can happen when a tier of the application is uninstrumented or uses a protocol Davis does not model. As of mid-2026, Dynatrace extends Davis under the "Dynatrace Intelligence" (hypermodal AI) umbrella, adding generative AI query interfaces (DQL-based) and predictive capabilities alongside the deterministic engine.

AppDynamics: Business-Transaction Correlation and Business Impact Analytics

AppDynamics organizes observability around the concept of business transactions: named units of work that map directly to user-facing operations (e.g., "checkout," "login," "payment-submit"). The Cognition Engine monitors the performance baseline of each business transaction and fires alerts when response time or error rate deviates. Business iQ extends this by correlating application performance with business KPIs injected via the Business Data API: revenue per minute, conversion rate, cart abandonment, and similar metrics.

This model is operationally well-suited to organizations where the observability audience includes business stakeholders, product managers, and NOC teams who need to understand application health in terms of revenue impact rather than infrastructure topology. The trade-off is that root-cause investigation still requires analyst effort. Business iQ surfaces the business impact immediately but does not automatically identify the causal infrastructure entity. Moving from "checkout latency is up 400ms and costing $X/minute" to "this specific database query on pod Y is the cause" requires the analyst to pivot into the APM flow maps and drill down manually, which is the twenty-minute investigation in the intro scenario.

Scope: Full-Stack Observability vs. APM and Business Analytics

Scope determines whether this is a platform consolidation decision or a point-tool decision. Both vendors claim "full-stack," but the architectures differ materially.

Dynatrace: Unified Platform Across Infra, APM, Logs, Digital Experience, and Security

Dynatrace covers infrastructure monitoring, APM, log management, digital experience monitoring (RUM and synthetic), runtime application security (vulnerability analytics, runtime application protection), and security posture management on a single platform with a shared data model and single UI. All signals are stored in the Grail data lakehouse and queryable via DQL (Dynatrace Query Language), which enables cross-signal correlation without manual data joining. A DQL query can combine APM traces, log lines, infrastructure metrics, and security findings in a single statement.

The breadth means teams can retire point tools (a separate SIEM for runtime vulnerabilities, a separate RUM platform, a separate log aggregator) and reduce integration overhead. The cost implication is that each capability tier is metered separately under DPS, so platform breadth does not automatically mean platform savings. Organizations that activate log analytics, RUM, security modules, and full-stack monitoring simultaneously should model all DPS line items before signing a commit volume, not just the APM tier.

AppDynamics: APM Core Plus Business iQ and Cisco FSO Integrations

AppDynamics is APM-first. Its native scope covers application performance monitoring, database monitoring, end-user monitoring (browser and mobile), and infrastructure visibility via the Machine Agent. Business iQ adds the business-impact analytics layer. Coverage beyond this, particularly network observability, cloud infrastructure cost analytics, and network topology, is addressed through Cisco Full-Stack Observability integrations: ThousandEyes for network intelligence, Intersight for infrastructure, and Cisco's broader observability portfolio.

For organizations already operating ThousandEyes and Intersight, the Cisco FSO bundle creates a legitimately integrated observability stack with cross-product correlation. For organizations that are not Cisco shops, assembling that same breadth requires independent purchasing decisions across multiple Cisco products, which reintroduces the integration and licensing complexity that a unified platform is meant to eliminate.

Pricing Model and Cost Predictability: DPS vs. Agent and CPU Licensing

Pricing model shape is the highest-anxiety topic for enterprise buyers on this decision, and both tools handle it differently. One publishes rates; the other does not.

Dynatrace Platform Subscription: Consumption-Based DPS With Published Rates

Dynatrace's DPS model uses a prepaid annual commitment in DPS units, which map to specific consumption rates published on the Dynatrace pricing page. As of mid-2026, the published rates are:

• Foundation and Discovery: $7/mo per host
• Infrastructure Monitoring: $29/mo per host
• Full-Stack Monitoring: $58/mo per 8 GiB host (billed at $0.01 per memory-GiB-hour)
• Kubernetes Platform Monitoring: $1.40/mo per pod
• Code Monitoring: $3.60/mo per container
• Log Analytics: $0.20/GiB ingest; retention billed as $0.0007/GiB-day (pay-per-query with $0.0035/GiB-scanned) or $0.02/GiB-day for bundled queries with 10-35 day retention
• Metrics: $0.15 per 100,000 datapoints
• Traces: $0.20/GiB (10 days retention included)
• Events: $0.20/GiB
• RUM: $2.25 per 1,000 sessions; $4.50 per 1,000 sessions with session replay
• Synthetic Browser Monitor: $4.50 per 1,000 actions
• HTTP Monitor: $1.00 per 1,000 requests
• Runtime Vulnerability Analytics: $13/mo per 8 GiB host
• Runtime Application Protection: $13/mo per 8 GiB host
• Security Posture Management: $5/mo per host

Dynatrace does not penalize for exceeding a committed DPS volume, and volume discounts apply at higher commitment tiers. The public rate card is a meaningful FinOps advantage: cost modeling is possible before the sales conversation begins, and consumption forecasts can be built from host counts, memory size, pod counts, expected log volume, and session counts using the published figures above.

AppDynamics Agent and CPU-Based Enterprise Licensing

AppDynamics pricing is quote-based across all tiers under Cisco ownership. Licensing is structured around agents: App Agents (per runtime instance or per CPU, depending on the license type), Machine Agents, Database Visibility Agents, Business Agents, and EUM licenses. The specific dollar rates per CPU or per agent are not published and are confirmed as unverified for this article. Buyers must engage Cisco sales or a Cisco partner to obtain pricing.

This structure creates two planning problems. First, cost modeling before the sales engagement is not possible with precision. Second, agent-based licensing means cost scales with the number of instrumented runtimes and machines, which in large Kubernetes environments can produce unexpected cost growth as pod counts scale dynamically. Cisco frequently bundles AppDynamics into broader Cisco FSO agreements, which can reduce per-unit cost for existing Cisco customers but also increases commercial dependency on the Cisco relationship.

What Each Model Means for Cost at Scale and FinOps

At scale, the cost composition of both tools diverges from the base APM rate. For Dynatrace DPS, the largest cost drivers beyond host-level monitoring are typically log ingestion and retention (log volumes in microservices estates can be substantial at $0.20/GiB ingest plus retention costs), RUM at session volume, and security modules if activated. A 500-host estate with active log analytics, RUM, and runtime security is materially more expensive than 500 hosts at the Full-Stack Monitoring base rate alone. FinOps teams should model each DPS line item against actual telemetry volumes, not just host counts.

For AppDynamics, the equivalent scale concern is the agent count multiplied by the per-CPU or per-agent negotiated rate, plus separate licensing for Database Visibility and EUM. Organizations with many short-lived containers or ephemeral workloads sometimes find that agent-based licensing scales less favorably than consumption-based models in cloud-native environments where instance counts are volatile. Because rates are not published, negotiation leverage and existing Cisco relationship depth significantly affect the actual TCO.

Cloud-Native and Kubernetes Support

Dynatrace deploys into Kubernetes via the Dynatrace Operator, which manages OneAgent DaemonSet deployment and offers CodeModules for injecting instrumentation at pod startup without modifying container images. The Kubernetes Platform Monitoring tier ($1.40/mo per pod) provides node, pod, and workload-level metrics, and Smartscape maps Kubernetes topology automatically, including service-to-service dependencies across namespaces. Service mesh environments (Istio, Linkerd) are supported, with Davis correlating mesh telemetry with application-level traces. Serverless coverage spans AWS Lambda, Azure Functions, and Google Cloud Functions via dedicated extensions.

AppDynamics supports Kubernetes via Helm chart deployment of the AppDynamics Cluster Agent, which collects pod and node metrics, and via App Agent injection for instrumented workloads. The Cisco FSO portfolio adds cloud-native integrations at the infrastructure layer through Intersight. AppDynamics' Kubernetes support has matured significantly since 2020, but the multi-agent model means that full-stack coverage of a polyglot Kubernetes cluster (Go services, Java services, Python services) still requires deploying and managing multiple agent types. Organizations building greenfield Kubernetes platforms frequently report lower operational overhead with the single-agent operator model.

Log Management and OpenTelemetry Lock-In

Log Ingest and Retention Economics

Dynatrace log management is built on the Grail data lakehouse and priced at $0.20/GiB ingest. Retention adds cost: $0.0007/GiB-day for the pay-per-query tier (where queries are additionally charged at $0.0035/GiB-scanned) or $0.02/GiB-day for the bundled-query tier with retention between 10 and 35 days. For high-volume log environments (thousands of microservices emitting structured logs), retention costs accumulate quickly and must be modeled explicitly. The pay-per-query tier is favorable for archives queried infrequently; the bundled-query tier suits environments with frequent log investigation workflows.

AppDynamics log analytics capability is available and has been enhanced under Cisco ownership, but specific ingest and retention pricing is not published and is confirmed as unverified for this article. Cisco's broader observability logging story also references integration with third-party log platforms, meaning some AppDynamics deployments route logs to a separate log management system rather than AppDynamics natively.

OpenTelemetry Support and Portability Implications

OpenTelemetry support is a proxy for lock-in risk. Dynatrace ingests OTel traces (OTLP), metrics, and logs natively into Grail. OTel data is processed alongside proprietary OneAgent signals and is queryable via DQL. This means an organization can instrument services with OTel SDKs and send telemetry to Dynatrace without deploying OneAgent, reducing the proprietary agent footprint. The data, however, is stored in Dynatrace's Grail lakehouse under proprietary retention and query economics, so signal portability coexists with data-layer dependency.

AppDynamics supports OTel for trace ingest, allowing OTLP-instrumented services to report to the AppDynamics Controller. However, the full depth of AppDynamics features, including business-transaction baselining, Business iQ correlation, and the Cognition Engine, depends on proprietary App Agent instrumentation. OTel-only instrumented services do not receive the same feature coverage as App Agent-instrumented services. This means that OTel adoption in an AppDynamics environment is additive at the edges but does not eliminate the proprietary agent dependency for core functionality.

For organizations with an explicit vendor-portability requirement, neither platform is OTel-native in the sense that migrating away is costless. Dynatrace's OTel support is broader and more deeply integrated, but both tools should be evaluated on the assumption that re-instrumentation cost is real if the platform changes.

Deployment Options: SaaS, On-Premises, Hybrid, and Air-Gapped

Dynatrace: SaaS-First With Managed Self-Hosted Option

Dynatrace is architected as a SaaS-first platform. The standard deployment is Dynatrace SaaS, where the cluster, Grail lakehouse, and Davis AI run in Dynatrace-managed infrastructure. Dynatrace Managed is a self-hosted deployment option where the Dynatrace cluster runs on customer-owned infrastructure, providing data-residency control. Air-gapped environments are addressed through the Dynatrace Managed option with restricted network configurations, but full feature parity with SaaS is not guaranteed in all air-gapped scenarios, and the operational burden of running Dynatrace Managed is substantially higher than SaaS. Customers considering Dynatrace Managed for compliance reasons should validate specific data-path requirements against current Dynatrace documentation before committing.

AppDynamics: On-Premises and Hybrid Flexibility for Regulated Environments

AppDynamics has a longer track record of on-premises and hybrid deployment. The AppDynamics Controller can be deployed on-premises or in a private cloud, and the on-prem deployment model has been a standard option since before the Cisco acquisition. This makes AppDynamics a more natural fit for environments with strict data-residency requirements, government or defense workloads, or air-gapped networks where SaaS connectivity is prohibited. Cisco's experience operating regulated environments at scale is an additional factor. Organizations in financial services, healthcare, or public sector with data-sovereignty constraints frequently find AppDynamics' on-prem option easier to justify to compliance teams than Dynatrace Managed.

Pros and Cons

It Depends: A Decision Framework by Org Size and Stack

Neither platform dominates across all deployment profiles. The decision is best framed by three variables: stack composition (cloud-native vs. hybrid/on-prem), vendor relationships (Cisco-aligned vs. vendor-neutral), and observability maturity (dedicated SRE/observability team vs. generalist ops).

Small and Mid-Size Cloud-Native Teams: Lean Dynatrace

Teams running predominantly cloud-native workloads on Kubernetes, with limited on-prem footprint and no existing Cisco commercial relationship, will find Dynatrace's operational model more efficient. OneAgent reduces agent lifecycle management to a single operator deployment. The published DPS rate card allows finance and engineering leadership to model costs from host counts and expected telemetry volumes before committing. Davis reduces the analyst overhead of root-cause investigation, which matters for small SRE teams managing large pod counts. The Grail lakehouse consolidates logs, metrics, traces, and security signals without a separate log aggregator, reducing integration cost for teams that lack dedicated tooling engineers.

Large Cisco-Aligned Enterprises: Lean AppDynamics

Organizations that have existing Cisco enterprise agreements covering ThousandEyes, Intersight, Catalyst, or Cisco's broader networking portfolio should evaluate AppDynamics inside the Cisco FSO bundle before treating this as a greenfield decision. The commercial leverage of bundling is real, and cross-product correlation between AppDynamics application telemetry and ThousandEyes network intelligence is an architecturally meaningful capability for organizations where application and network observability are managed together. Business iQ is also a genuine differentiator for enterprises where observability must surface directly to business stakeholders, not just engineering teams.

Regulated, Hybrid, or Air-Gapped Environments: Weigh AppDynamics On-Premises

Government, defense, financial services, and healthcare organizations with data-sovereignty mandates or air-gapped network requirements should evaluate AppDynamics on-premises more seriously than the typical cloud-native SaaS comparison suggests. AppDynamics has a longer and more mature on-prem deployment track record. Dynatrace Managed is an option but carries higher operational burden and potential feature gaps in fully air-gapped configurations. Before choosing Dynatrace in a regulated environment, validate the full feature set available in Dynatrace Managed against your compliance requirements, and factor in the engineering cost of operating the managed cluster.

Final Verdict

The dynatrace vs appdynamics decision does not have a universal answer, and any comparison that declares a flat winner is either ignoring deployment constraints or vendor-aligned. The architectural differences are real and consequential.

Dynatrace wins on instrumentation simplicity, AI-driven root-cause automation, pricing transparency, and cloud-native Kubernetes maturity. For organizations prioritizing operational efficiency of the observability platform itself, those properties carry significant weight over a multi-year contract. The consumption-based DPS model rewards FinOps discipline and allows granular cost allocation by capability. The published rate card is a genuine differentiator for procurement teams who need cost modeling before a sales engagement.

AppDynamics wins on business-impact analytics depth, on-premises and hybrid deployment maturity, and commercial leverage for Cisco-aligned enterprises. The Cisco FSO bundle is not a marketing convenience for organizations already running ThousandEyes and Intersight; it is an integration architecture with real operational value. Business iQ's ability to translate application telemetry into business KPIs remains a differentiated capability that Dynatrace's platform does not replicate directly.

Decision checklist before signing either contract:

• Model Dynatrace DPS costs across all intended capability tiers (host monitoring, log analytics, RUM, security modules) using the published rate card, not just the base APM line
• Request AppDynamics pricing broken down by agent type and count for your actual estate; model for 20-30% pod-count growth if running Kubernetes
• Determine whether your organization has an active Cisco enterprise agreement and what bundle pricing is available
• Assess data-residency and air-gap requirements before assuming Dynatrace SaaS is viable
• Evaluate whether your primary observability audience is engineering (favors Davis RCA) or business stakeholders (favors Business iQ)
• If the decision is close, run a parallel POC with both tools against a representative workload tier; instrument the same set of services with each platform and compare alert quality, RCA speed, and operational overhead directly

Verdict: DRAW. The correct platform is determined by stack, vendor relationship, and deployment constraints. Neither tool is categorically superior across all enterprise contexts.

Frequently Asked Questions

Is Dynatrace more expensive than AppDynamics?

It depends on the estate. Dynatrace publishes a public DPS rate card, so cost is modelable before the sales engagement. AppDynamics is quote-based with no published rates. At scale, Dynatrace costs are driven by host memory size, log volume, and activated capability modules. AppDynamics costs scale with agent count and CPU count. Neither is universally cheaper; total cost depends heavily on estate composition and negotiated rates.

Can Dynatrace replace AppDynamics, and what does migration cost?

Dynatrace can cover the same APM and infrastructure scope, but migration is not a swap. App Agents must be replaced with OneAgent or OTel instrumentation across every host and container. Business-transaction definitions, alert policies, SLOs, and dashboards in AppDynamics do not have direct import paths into Dynatrace and must be rebuilt. Organizations should budget for a parallel-run period and significant re-instrumentation and dashboard-rebuild effort before treating DPS pricing as the total migration cost.

Which is better for Kubernetes?

Dynatrace is operationally more efficient for Kubernetes-heavy estates. The Dynatrace Operator deploys OneAgent as a DaemonSet or via CSI injection, covers dynamic pod counts without manual agent reconfiguration, and offers a per-pod billing tier at $1.40/mo per pod. AppDynamics supports Kubernetes via the Cluster Agent and App Agent injection but requires managing multiple agent types for polyglot workloads, increasing operational overhead in large clusters.

Does AppDynamics support OpenTelemetry?

AppDynamics accepts OTel trace data via OTLP, allowing OTel-instrumented services to report to the Controller. However, core AppDynamics features, including business-transaction baselining and Business iQ correlation, require proprietary App Agent instrumentation. OTel-only instrumented services receive reduced feature coverage. Dynatrace ingests OTel traces, metrics, and logs natively into Grail alongside OneAgent signals, with broader OTel integration depth.

What is the difference between Davis AI and Business iQ?

Davis is Dynatrace's deterministic causal AI engine. It traverses Smartscape's real-time topology graph to isolate the root-cause entity automatically, producing a single hypothesis rather than a ranked list. Business iQ is AppDynamics's business-impact analytics layer. It correlates application performance degradation with business KPIs (revenue, conversion, session counts) to quantify the business cost of a problem. Davis answers "what broke and why"; Business iQ answers "what does this degradation cost the business."

Is AppDynamics still a strong choice under Cisco?

AppDynamics remains a capable platform under Cisco ownership, with continued investment in the Cisco Full-Stack Observability portfolio. The legitimate due-diligence question is where AppDynamics sits in Cisco's product roadmap priority relative to ThousandEyes and Cisco's broader networking observability strategy. Organizations should ask Cisco directly for a 24-month product roadmap commitment and evaluate how deeply AppDynamics-specific capabilities (not just FSO bundle features) are being invested in before signing a multi-year agreement.

Which offers on-premises or air-gapped deployment?

AppDynamics has a longer and more mature on-premises deployment track record. The AppDynamics Controller can be deployed on customer infrastructure, which suits air-gapped, data-sovereign, and regulated environments. Dynatrace offers Dynatrace Managed for self-hosted deployment, but it adds operational complexity and may not achieve full feature parity with Dynatrace SaaS in all air-gapped configurations. Organizations with strict data-residency or disconnected-network requirements should validate Dynatrace Managed against their specific requirements before ruling out AppDynamics on-premises.