Edge Delta vs Fabrix.ai (2026)

About twenty event connectors cover Argo CD, Atlassian, AWS, CircleCI, Databricks, GitHub, GitLab, Grafana, Harness, Jenkins, LaunchDarkly, Linear, Microsoft Teams, PagerDuty, Sentry, Slack and more, about twenty-six streaming connectors cover Kubernetes, Kafka, OTLP, Syslog, SNMP, Splunk, CrowdStrike FDR, Windows Events and cloud object stores, and a destinations library reaches some forty systems across observability tools, data lakes and SIEMs.

Event connectors work in both directions, receiving events from external systems and taking outbound actions such as GitHub pull request operations, Jira issues, PagerDuty incidents and Slack and Teams messages, written directly into those systems. Custom remote MCP servers and a Custom API source let customers add their own connections, and the connector classes span CI/CD, incident management, cloud, collaboration, data platforms and security.

Sourcedocs.edgedelta.com/ai-connectors, /event-connectors and /destinationsread 2026-09-08

Agents work today with roughly thirty named commercial products, covering observability and monitoring (Dynatrace, New Relic, Datadog, SolarWinds, PRTG, Zabbix, Prometheus, Grafana), ITSM (ServiceNow, Jira Service Management), network and virtualization (Cisco Meraki, Catalyst Center, Nexus Dashboard, Cisco Firepower, Juniper Mist, Infoblox, VMware vCenter), security and logs (Splunk, Elastic, IBM QRadar), storage and infrastructure (NetApp, BMC, HPE) and cloud (AWS, Azure, GCP), with Dell Technologies, ScienceLogic and PagerDuty as well.

Fabrix.ai lists Cisco, Splunk, IBM, AWS and NVIDIA as partners. The Automation Fabric opens tickets, sends notifications and runs remediations, agents route safe remediation and audit and act on network configuration, and the orchestration layer connects agents, tools, APIs, data sources and workflows. In Fabrix.ai's words, "Your tools don't need to wait for their vendors to ship MCP support.

Fabrix.ai dynamically generates MCP tools for any system," modern or legacy, with zero code. A Universal Connector, guides for integrating data sources and a browsable catalog of 224 bot extensions round it out. Fabrix.ai does not publish a matrix of supported versions.

Sourcefabrix.ai homepage Agents for Any Tool tab and partner pages, docs.fabrix.ai extension catalogread 2026-09-14

Two mechanisms coordinate work. OnCall AI is the orchestrating teammate that routes requests to specialists, coordinates workflows across several teammates and synthesizes findings, working across a set of role-specific teammates with de-duplication and handoffs.

The Workflows builder is a node graph with Start, Teammate, Task, If/Else, Transform, Action and Note nodes, where Action nodes cover email, Slack, Jira, Microsoft Teams, PagerDuty and starting an AI conversation, and Edge Delta ships Default Workflows and a Workflow Patterns library. Branching, transformation and conditional routing run across both agents and external systems, and Monitors can invoke workflows.

Sourcedocs.edgedelta.com/workflows-section and /glossaryread 2026-09-08

Agent Orchestrator is Layer 02 of the platform, "a no-code orchestration layer that connects agents, tools, APIs, data sources, and workflows," running single agents or handing work between agents across four operational domains, from ITOps and NetOps to SecOps and BizOps. No code workflows run "with conditions, approvals, retries, rollback," and the observability layer shows every branch and retry in its traces, along with fallbacks and loops.

Customers build and run agentic workflows by drag and drop with a built in task library, with an Agent Builder for agents and execution flows and a path from prompt to production workflow. The Build Your Own Agent guide sets out the steps, choosing a template and defining the operational intent, shaping responses with the prompt builder and attaching data sources and thresholds, then going on to "configure constraints, escalation logic, or feedback loops."

Agents can also call one another, since they "discover and collaborate with other Fabrix.ai or third-party agents" and "participate in multi-agent workflows across network, app, and service domains," and Fabrix.ai is a member of the AGNTCY collective. Fabrix.ai does not publish versioning or concurrency behavior.

Sourcefabrix.ai homepage Agent Orchestrator layer, fabrix.ai/agenticai orchestration block and BYOA guideread 2026-09-14

A monitor alert, a PagerDuty incident or an event connector starts an investigation thread with no acknowledgment step in between.

Threshold, change, anomaly, pattern-anomaly, composite and synthetic monitors can each trigger an AI Team workflow or post into an AI Team channel, pipeline triggers fire at the edge for lowest latency, and about twenty event connectors listen to CI/CD, incident and code-hosting systems. Streaming connectors add inbound queues and listeners including Kafka, Azure Event Hub, Pub/Sub, SNMP Trap, Syslog and HTTP.

People reach the AI Team through its Channels and Direct Messages, Slack, Microsoft Teams, email, PagerDuty and webhooks, and events, schedules and inbound queues all start work.

Sourcedocs.edgedelta.com/monitors, /send-to-workflow and /event-connectorsread 2026-09-08

The Data Fabric ships Telemetry Pipelines, a Universal Connector and Pipeline Studio and supports streaming and batch ingestion for hybrid IT environments, and Fabrix.ai treats event driven architecture, with real time streaming and orchestration, as a required capability. Each agent family acts in real time.

NetOps agents "detect config drift and policy violations the moment they happen", observability agents "identify emerging issues and forecast failures before they impact services" and "correlate and suppress alert noise before it reaches your team", and asset agents keep asset records current. Agents also trigger each other, with action nodes invoking other agents, agents that "discover and collaborate with other Fabrix.ai or third-party agents," and work passing between agents across the four operational domains. Agents deploy on a schedule or on an event trigger. Fabrix.ai does not describe a webhook endpoint, cron syntax, an event subscription model or an inbound API trigger.

Sourcefabrix.ai homepage NetOps and observability agent benefits, fabrix.ai/agenticai data fabric and multi-agent rowsread 2026-09-14

A Knowledge Graph is a core AI Team component, Knowledge Libraries are configured at pipeline level, and an indexed back end holds logs, metrics, patterns, traces and events that the customer queries through a published Common Query Language.

Around it sit a Log Inventory, Metrics Inventory, Patterns and Anomalies views, a Service Map, log-to-trace correlation, and Rehydrations that pull archived data back into the index. Teammates search this structure through MCP tools for log, metric, trace and event search, and it persists, scales past the context window and stays queryable. Telemetry Pipelines shape data before it reaches a model.

Sourcedocs.edgedelta.com/knowledge-graph, /knowledge-library and /cqlread 2026-09-08

Living Ontology, one of the platform's three foundations, is Layer 03, "a continuously updated enterprise knowledge graph that grounds every agent in real assets, topology, dependencies, events, services, owners, and business context," in place of stale snapshots or disconnected alerts, a live graph that keeps mapping every entity, dependency and topology across the IT environment.

It is shared across agents as "one shared intelligence layer," with "full-context reasoning across every team and domain," and Enterprise Ontology is a component of the Agent Orchestrator. The Data Fabric feeds it, providing "rich and automated data enrichment with service topology and business context," with Universal Connector, Telemetry Pipelines, Pipeline Studio and Data Discovery & Enrichment among its named modules and 224 bot extensions.

Domain agents work together over a shared semantic model. Every agent is "grounded in your environment" rather than guessing, and agents working with Splunk "enrich Splunk data with full IT context from the enterprise knowledge graph." Fabrix.ai does not describe a retrieval mechanism, embedding approach or query interface for the graph, though it has its own query language, CFXQL.

Sourcefabrix.ai homepage Living Ontology layer and section, fabrix.ai/agenticai data context rowread 2026-09-14

Teammates keep memories from prior work in two scopes. Memories shared across the organization draw on previous analysis findings, so a similar pattern investigated before speeds up the current investigation, and personal memories hold an individual user's preferred analysis approaches and prior decisions.

Each can be toggled independently, retention policies are set in settings, and turning memories off or expiring them destroys no business record. What teammates learn about normal behavior across services, and historical baselines, sit in the anomaly detection model. Separately, the sandbox workspace persists across sub-steps within one investigation, and threads retain full history for handoffs.

Sourcedocs.edgedelta.com/ai-team-sandbox and /ai-team-settingsread 2026-09-08

The Living Ontology describes the customer's estate. Agents "learn context, refine thresholds dynamically, and use feedback to improve detection," and reconfigure suppression logic on their own, which changes a model or a configuration. Smart Context Management offers caching and retrieval for large datasets with savings in tokens and latency, "a high-speed intelligent cache that supplies your AI agents by overcoming context window limitations." It exists to save tokens, and the control plane reports cache savings as a cost metric.

Within a session, Agent-0, the Copilot, works in an iterative loop, "Prototype in Copilot, iterate, then simply ask to create agent with persona, tools, prompts, and workflow auto-packaged," and copilots "allow teams to validate responses before agents are operationalized." Fabrix.ai does not describe a memory layer, a retention period, expiry or purge path, a memory object per user or team, or an agent carrying state from one run into the next.

Sourcefabrix.ai/agenticai smart context and prompt-to-agent tabs, fabrix.ai homepage cost insights cache savingsread 2026-09-14

Teammates are read-only by default and ask before any write action, with restarts and rollbacks named as the kind of operation that needs approval. On the Issue Details page a user approves or denies AI actions and verifies or dismisses the steps they own. Guardrails add a configurable layer with Domain Overrides, a Data Boundary controlling what reaches the model, and a Posture Report, and code execution inside an investigation follows the same permission controls as every other teammate operation. Connector scopes and rate limits set per agent bound what each teammate can do.

Sourcedocs.edgedelta.com/issue-details, /guardrails and /ai-team-securityread 2026-09-08

The AIOps agents promise Human-in-the-loop safety, "Your team approves every action," with transparency and control at every step, and for cybersecurity "Every automated action governed by guardrails, explainability, and human approval." Approval steps are part of the orchestration layer, where no code workflows run "with conditions, approvals, retries, rollback," so a customer composes approval steps into a workflow the way they compose conditions.

The Agent Control Plane is "a shared control layer for observability, approvals, policy enforcement, audit trails, PII masking, model usage, token spend, and agent behavior across teams, tools, and LLM providers." Agents "can run autonomously or with approval workflows in place for sensitive actions," and the customer decides per workflow, and rollback lets a team undo an action as well. Fabrix.ai does not describe an approval queue screen, an approver role model, an escalation path or a record of who approved.

Sourcefabrix.ai homepage AIOps and cybersecurity benefit blocks, control plane layer, fabrix.ai/agenticai orchestration building blockread 2026-09-14

SAML handles authentication and authorization, with an implementation guide and six identity providers covered individually (Azure AD, Okta, Cisco Duo, Google, OneLogin, Ping Identity).

Administrators manage Users and Groups and API tokens, role-based access control extends to RBAC scoped to Packs, each teammate has its own connector scopes and rate limits, and cloud identity integrates through AWS IAM roles and service accounts, AWS assumed roles, GCP authentication and Kerberos. Secrets management with master-key encryption sits underneath.

Edge Delta holds SOC 2 Type II, and its trust center at trust.edgedelta.com covers data handling and AI model provider policies, alongside a Strengthening Security and Compliance page.

Sourcedocs.edgedelta.com/saml-integration, /users, /api-tokens and trust.edgedelta.comread 2026-09-08

Fabrix.ai states it is SOC 2 certified, with "Role-based access, full audit trails, and guardrails on every agent interaction," and keeps a security overview page at /security/. Role based access reaches down to each agent, since AI Personas provide "RBAC for models, tools, prompts," a scoping that "presents only persona-relevant MCP tools and data to LLM," so a customer can set which tools and data each persona reaches.

The agentic platform adds "a built-in policy engine, role-based access, and audit trails, with enterprise-grade guardrails to restrict agent permissions, scope of actions, escalation pathways." Fabrix.ai's vX release lists single sign on (SSO) among its enterprise foundation capabilities, and data masking and PII redaction round it out. Fabrix.ai does not name the SOC 2 type or auditor, and does not publish a report date, scope or SCIM detail.

Sourcefabrix.ai homepage security block and footer badge, fabrix.ai/agenticai personas and governance rowsread 2026-09-14

Audit Logs under Administration record Edge Delta activity, and the platform's own audit logs can be ingested back through a pipeline, so the record can be exported into the customer's SIEM. Around it sit an Activity view in the AI Team, AI Team Events, per-thread detail showing what the agent did and what it consumed, an AI Overview dashboard covering live AI Team activity, and Issue Details, where each recommended step can be verified or dismissed by its owner. Every agent action is logged, so an investigation can be reconstructed afterward.

Sourcedocs.edgedelta.com/audit-logs, /ingest-audit-logs and /activityread 2026-09-08

A distinct Agent Control Plane records the agents' own runs, with "every AI action traced, governed, and explainable." "Every agent run leaves a complete trace," from the first prompt to the final action, showing every tool call, every model invocation, every branch and retry, with full payload visibility and PII masking built in.

The AI Interaction Flow Tracer lays out the chain from persona and prompt through context and tools to the model and result, with a payload view with redaction, retries, fallbacks and loops, a latency breakdown per step and exportable traces.

AI Explainability promises to "make every AI decision transparent, auditable, and defensible," and each run includes a decision trace of the reasoning chain and tool calls, with evidence and policy checks, "so you can explain any outcome to any stakeholder," and run metadata on model and version, MCP tools, and persona and scopes.

Admins get one dashboard for AI across teams, apps and providers, with drilldowns from organization to team to run, leaderboards by user, agent and persona, and reliability metrics such as success rate and failed run cost. Fabrix.ai does not publish a retention period for traces.

Sourcefabrix.ai homepage Agent Control Plane tabs, flow tracer and explainability panelsread 2026-09-14

The pipeline and processing tier installs into the customer's own infrastructure, with paths for Kubernetes (Helm, kubectl, manifests, OpenShift, SELinux-enforced clusters), Linux, Docker, Amazon ECS, macOS and Windows, plus a Terraform provider, GitOps, multi-tenant and multi-cluster patterns, in-cluster processing, in-cluster destinations, a Local Storage destination and a Guardrails Data Boundary.

That gives the customer control over where telemetry is processed and what leaves its estate. The AI Team itself runs its control plane in Edge Delta's cloud, where the sandbox for each investigation is provisioned, and it cannot be self-hosted, has no region list and offers no option to run in the customer's environment.

Sourcedocs.edgedelta.com deployment patterns, in-cluster processing and ai-team-sandboxread 2026-09-08

The platform is "SOC 2 certified, deployable on-prem or in your private cloud," one agentic platform to design, deploy, operate and observe AI agents, on prem or in the cloud, so the software runs inside the customer's boundary.

Models can run there too, with "on-premise / self-hosted LLMs" deploying open source models within the customer's secure infrastructure, "best for data sovereignty, compliance, or low-latency edge scenarios," so prompts need not leave the customer's network. Agents "run natively on your NVIDIA GPUs," on bare metal, private cloud or hosted endpoints.

Fabrix.ai offers a Sovereign AI use case and installation guides for RDA Fabric and its components. It does not publish a region list or residency commitment for the hosted option.

Sourcefabrix.ai homepage security and deployment blocks, fabrix.ai/agenticai multi-LLM tab and quick start, docs.fabrix.ai installation guidesread 2026-09-14

Six specialized teammates ship out of the box, AI SRE, Security Engineer, Software Engineer, Work Tracker, DevOps Engineer and Cloud Engineer, each with its own production-ready system prompt, connector scopes and channel access. Separately, Packs are a product feature with a browsable Packs Catalog of roughly forty-three entries (Akamai, ArgoCD, Auth0, AWS CloudTrail, Cisco ASA, Cloudflare, Fortigate, Istio, Nginx, Okta, Palo Alto, Splunk Rosetta Stone, ZScaler and the rest), a Packs Library, and flows to create a pack, create one from an existing pipeline, add one to a pipeline, deploy changes to active packs and configure RBAC per pack.

Sourcedocs.edgedelta.com/pack-catalogue and /specialized-teammatesread 2026-09-08

A browsable Agent Catalog lists the prebuilt agents at /agenticai/agent-catalog/ and /ai-agents/. Eight named families, Agent-0 (Copilot), Digital SRE / AIOps, Observability Agents, NetOps Agents, ServiceOps Agents, CollabOps, SecOps Agents and BizOps Agents, hold roughly thirty individually named agents.

Operations work has the Asset Health Reporting Agent, Uptime & Availability Agent, Anomaly Detection Agent, KPI Forecasting Agent, Alert Optimization Advisor, RCA Agent, Incident Assignment Agent, Remediation Agent and DEX Analyst. Network and asset work is covered by the Config Compliance Agent, ACL Change Audit Agent, Netflow Analysis Agent, BGP & Routing Intelligence Agent, SACM Asset Intelligence Analyst and Asset Lifecycle & Capacity Analyst.

For Splunk and the service desk there are the Splunk SIEM Agent, ITSI Analyst & Resilience Agent and Service Desk Agent, and the security set runs to the Automated Reconnaissance Agent (VAPT), Exploit Assistant, User Behavior Analysis Agent, Patch Prioritization Agent, Compliance Mapping Agent and Control Validation Agent. Fabrix.ai suggests adopting them a team at a time, "Start with one team, then scale across ITOps, NetOps, SecOps, SRE, and BizOps."

The Quick Start suggests customers "onboard out-of-the-box agents" to speed deployment, using prebuilt agents for common telecom and enterprise use cases such as RCA, Anomaly Detection and Service Migration, and "use out-of-the-box agents, customize them for your workflows, or build your own." Solution Packs appear separately under the Automation Fabric, and the 224 bot extensions are integration building blocks.

Sourcefabrix.ai homepage agent tabs and family lists, fabrix.ai/agenticai products navigation and quick startread 2026-09-14

Teammates can run on Anthropic, OpenAI or Google models. The platform is model-agnostic across all three, so administrators can switch models, attach a distinct foundation model to each teammate to match model to role, and stay current through a maintained model catalog. The trust center lists policies for multiple AI model providers. AWS is the cloud host, not an additional model provider.

Sourcedocs.edgedelta.com/teammates, held from 2026-07-08 launch and documentation basis, re-stampedread 2026-09-08

Multi LLM Choice leads the platform's capabilities, "Support featured LLMs. On-prem or cloud. Seamless integration," and is NVIDIA ready. Fabrix.ai calls it a differentiator that "Supports integration with multiple LLMs (e.g., OpenAI, Anthropic, open-source models) to enable agent flexibility, model fallback, use-case-specific tuning, and vendor-neutral AI orchestration."

The Quick Start has the customer "select your preferred AI deployment model" from two options, cloud LLMs from OpenAI, Anthropic or other public APIs, or on premises or self hosted LLMs, with open source models such as LLaMA or Mistral running within the customer's secure infrastructure, adding that "Fabrix.ai supports hybrid and multi-LLM architectures." Customers can bring their own model and host it themselves.

The control plane reports the mix of providers and models, with share and trends, and cost by LLM, user, persona and agent, so a customer can see which model each agent uses and what it costs. AI Personas act as "RBAC for models, tools, prompts," letting admins control which models a role may reach, and Evaluators lets a customer "run side-by-side model evaluations on your own data and rank quality, cost, and reliability before you deploy." Fabrix.ai does not publish a model list, version matrix or default.

Sourcefabrix.ai/agenticai multi-LLM tab, building blocks and Quick Start, fabrix.ai homepage control plane model mixread 2026-09-14

A public REST API at api.edgedelta.com/v1 has an interactive Swagger reference, API token and organization ID setup, endpoints to retrieve, update, validate and deploy pipeline configurations and to query metric timeseries for external integrations, an error table, and a complete OpenAPI 3.0 specification available for download at api2.edgedelta.com/swagger/doc.json. Alongside it are an edx CLI with a full command reference, a Terraform provider, and Deputy-style embedding through Cloud Pipelines.

The Edge Delta MCP server, published as Connect AI Agents to Edge Delta, lets an outside agent drive the platform, with tools for log, metric, trace and event search, pipeline management, live capture and processor dry runs. It is separate from the outbound custom remote MCP connector, which lets teammates use other systems, and from the internal MCP connector that bridges teammates to Edge Delta's own data.

Sourcedocs.edgedelta.com/api and /edge-delta-mcp-serverread 2026-09-08

Fabrix.ai offers a RESTful API, served as a Swagger interface at /beginners_guide/swagger_api/, and a Command Line Interface guide at /beginners_guide/sdk/, both public on its documentation site at docs.fabrix.ai. The developer material also covers architecture, installation and data ingestion guides, CFXQL, a named query language with its own reference, Grok pattern references, and a catalog of 224 bot extensions browsable by name, among other references and examples.

The platform also runs an inbound MCP server. "Fabrix.ai includes a powerful MCP server that exposes data pipelines and enriched telemetry, automation workflows and runbooks," which "allows external agents, tools, or orchestration layers to seamlessly access and interact with Fabrix.ai's real-time intelligence." Universal MCP Server & Tooling is a component of the Agent Orchestrator.

Sourcedocs.fabrix.ai index, RESTful API and CLI guides, fabrix.ai/agenticai MCP server descriptionread 2026-09-14

A Performance page tracks teammate scores, response times and token usage per thread. The AI Team Sandbox, despite its name, is not for evaluation; it is a secure isolated code-execution environment OnCall AI uses during investigations, spinning up when tool results exceed roughly 10,000 characters so the agent can write Python or Bash to analyze data it could not otherwise fit in context.

The pipeline release path gates changes through validate, save, version and deploy, with /confs/validate returning a valid flag and reason, which confirms a configuration installed correctly. Teammate Version History records what changed without scoring it, and Edge Delta does not describe scored test cases, a pre-deployment harness or a controlled optimization loop for teammate behavior.

Sourcedocs.edgedelta.com ai-team-performance, ai-team-sandbox and api-exampleread 2026-09-08

Evaluators sits in the Agent Control Plane beside Agent Observability and Cost & Token Insights, and its message is "Don't guess which model is best for your use case," and it lets a customer "run side-by-side evaluations on your own data and rank quality, cost, and reliability before you deploy."

The agent lifecycle runs through review, a dry run that simulates without executing and a test on sample data for limited iterations, then deployment and decommissioning.

Evaluations include A/B/C tests per use case, metrics for accuracy and factuality, coherence and safety, operational measures of tool calls, latency and tokens, and cost per result, human ratings and scoring against ground truth, and leaderboards, recommendations and audit reports, all run by the customer per use case on its own data.

In production, the observability layer reports quality, accuracy and coherence metrics for every run, plus success rate and failed run cost, and the Quick Start closes with testing on synthetic or historical data and iterating on prompts, thresholds or workflows based on real time performance. Fabrix.ai does not publish its methodology or regression suite detail.

Sourcefabrix.ai homepage Evaluators panel and control plane, fabrix.ai/agenticai Deploy & Test quick startread 2026-09-14

The closest thing to computer use is the AI Team Sandbox, which gives every investigation an isolated code interpreter with file system access, bash commands, the ability to write, compile and run Python or Bash, local storage for large tool outputs, and Git repository cloning to analyze full codebases, and which is deprovisioned when the investigation ends. The Exec source node runs scripts and commands to collect logs. Edge Delta offers no hosted or local browser, desktop session or remote computer control, and teammates reach outside systems through connectors and APIs instead.

Sourcedocs.edgedelta.com/ai-team-sandbox and /exec-input-noderead 2026-09-08

Fabrix.ai does not describe a hosted or local browser, a desktop session or remote computer control that the agent drives. Agents act programmatically. The Automation Fabric runs workflows and remediations, the orchestration layer "connects agents, tools, APIs, data sources, and workflows," and integrations reach roughly thirty named systems through connectors, bots and generated MCP tools.

Fabrix.ai "dynamically generates MCP tools for any system," modern or legacy, "API-driven or not," with zero code, from modern APIs to decades old legacy systems, but it does not say how it reaches a system with no API, and no screen, browser, cursor, UI element or visual model is named as a target. Streaming telemetry collection feeds the knowledge graph.

Sourcefabrix.ai homepage Agents for Any Tool tab, fabrix.ai/agenticai MCP tools block, docs.fabrix.ai extension catalogread 2026-09-14