Agent Ops

the control tower for an enterprise’s agent estate it governs what agents can do, observes what they are doing, and ensures resilience, compliance, and cost control across the full lifecycle innovation semantic touter routing relevance with precision bayesian react reacting with reason and bayesian guardrails nlq for operational searches pydantic templates prevent prompt injection attacks value proposition accelerates time to market for agentic solutions — teams can create, reuse, and scale workflows without technical expertise enables secure, rapid onboarding with mcp based integrations, governed access, and an ecosystem that connects enterprise and approved public knowledge reduces engineering overhead by removing the need to build and maintain orchestration infrastructure drives ai adoption across delivery, consulting, and solutioning — unlocking faster insights, automation, and decision making boosts bridge platform adoption across advanced delivery, dws, cloud, and bridge app delivers enterprise grade governance with built in ci/cd automation, tool integrations, memory management, and cost monitoring prevents reinvention by enabling teams to reuse ai workflows across projects and domains positions bridge as the enterprise ai backbone — scalable, extensible, and ready for future multi agent innovation agents ops dashboard the agent library contains your active agents, making effective management essential to support this, an observability capability is needed to monitor and manage agents while addressing security and compliance to access the dashboard, go to services > platform services > knowledge & ai > agents ops the dashboard provides key monitoring insights to enhance operational efficiency you can access to agent specific key performance indicators (kpis), such as the total number of executing agents, metrics on successful and failed executions, and those active for long durations you can assess tenant health, deployment status, total execution counts, and trends in token usage the dashboard includes an agent topology section displaying model usage, agent status, and a complete agent list overall tenant health the report shows whether all components are functioning properly this overview covers a specific time frame, and users can refresh the information to see the latest updates total agents registered this section shows the total number of agents registered during the specified period alongside the count, a percentage change indicates whether registrations increased or decreased compared to the previous period unique deployments here, you'll find the total number of unique deployments that have taken place additionally, a percentage change shows any increase or decrease in deployments relative to the last reporting period total executions this metric shows the cumulative number of executions during the defined timeframe, along with a percentage change indicating whether this figure increased or decreased compared to the previous period successful executions this section lists the total number of successful executions it also provides a percentage change to show whether successful outcomes increased or decreased compared to earlier results failed executions this section reports the total number of executions that did not succeed for context, a percentage change shows how this figure compares—whether it has increased or decreased—relative to the last reporting period token usage this metric captures total token usage, including both input and output tokens this figure is accompanied by a percentage change comparing current usage to the previous period, highlighting any upward or downward trends weekly execution analysis (7 day overview) this comprehensive report offers an in depth examination of successful and failed executions over the past week it aims to identify trends and discern patterns within this specific timeframe, providing valuable insights into performance and operational efficiency agents distribution metrics for each agent encompass execution counts, success rates, and recent statistics, along with their status (active, inactive, or in draft mode) and overall capacity insights model in use this section outlines the models currently in operation, including those scheduled to be activated soon agent status a visual representation (in the form of a donut chart) illustrates the distribution of agents based on their current status, offering a quick snapshot of their operational condition agent type this breakdown categorizes agents by type, showing the number of agents in each category agent topology a graphical display depicts the arrangement and connections between agents, highlighting the overall architecture of the agent network active agents table this table lists active agents and details essential metrics such as agent name, total executions, success rate, and the date of the last execution mcp server and tools registry table this section provides an overview of the mcp server and the total count of available tools, helping track resources and tool utilization efficiently