Good morning everyone. I'm PK, and today we'll explore how an event driven observability layer can finally light up the blind spots between ECA and the MA teams. Think of this session as journey from today's file. Drop headaches to your future where design changes surface in real time. Before. They burn either schedule or budget or both. Here is map. We'll begin with collaboration gap. Why late found issues are still the norm? Then we'll find defined that event driven vision. We'll unpack three core building blocks named Schema audit. And library sync and two cross cutting enablers, taxonomy and resilience. A quick architecture walkthrough shows how it all fits. We look at pilot projects, technical and business KPIs. Finally, we'll touch on monitoring roadmap. By the end, you will see exactly where to start. Your own tool chain. Let's ground the problem even in 2025. Most hardware programs pass files between ECA and MA at a gated milestones. A study across eight boards found over 10 snapshots per spin. Each snapshot a potential rerouting. Industry data shows that late reworking inflates cost eight to 10% and drag schedules four to six weeks. People end up in detective mode, which file version, which variant who change the key part when that fog sets in. Design energy shifts from. Innovation to forensic triage. That's the gap we must close. Hope you all agree with me. Here's a story you'll recognize. Mid program mechanical changes at chais rib by three millimeters. Yeah, it's just three millimeters. Keyboard strengths. But ECA keeps routing blissfully unaware. A week later, the ya screams and suddenly you are starting at a 250 K wrist pin and a six week slip as a sample. Numbers root cause isnt bad tools. It's timing of information, isn't it? The insight actually arrived too late to matter. So why events first, they are push-based. A tool emits a lightweight message. The instance something meaningful happens. Second events or context reach. Bundle geometry, hashes, user IDs, variant tax. Third context leads downstreams check file automatically, and third even flows through a broker that is resilient. If M card is done, ECA can keep publishing the system. Replace. Once MCA is back, the bottom line is. Decision that used to wait. This now happens in minutes. Our first billing block is a shared schema. Every ECA change or MCA update follows the same contract item id, revision id, geometry, hash user. Timestamp. Optional blocks. Capture CC variance for bomb options. Tools. Validate locally mal informed events, never leave the workstation. Consistent structure is what lets hundreds of tools speak one language without. Back at the code Second, audit, logging every event also lands in an immutable store. We track general workflow, security, file access logs. Who changed what and when These logs. Cut your cost time from hours, two minutes and satisfy the traceability mandate in medical and automation automotive hardware. They are exportable to CSV for compliance reviews. No more screenshot archeology, and it's library sync. Third library synchronization. Using ECA ed Library xml, we run export library, sync Library, and multi revision check-ins that keeps symbols and footprints aligned across tools like OrCAD, Allegro, and nx, where they live. The status file success flag becomes another event. So analytics can spot its tail footprint before it sneaks into schematic. Now we layer two crosscut enablers, geometry events, board outline, key ports, mounting holes. Component events like placement moves, part swaps analysis, events like DFM or thermal rule results during the projects. Lightweight heuristics, nothing exotic, mostly DUM, and the bonds cut irrelevant. Charter by 54%. That means engineers only see the signal, not the noise. Of course, events are great until something starts so we embed back pressure and replay. If MCAT goes offline, the broker queues safely and replace on recovery. Compensating transactions, rollback partial rights, A checksum, verifier blocks, corrupt payloads, and a circuit breaker isolate flaky. PLM APIs. The net result observability you can actually trust at scale. Putting it together, ECA and MA, published into a Kafka broker stream processor, compare events, geometry, deltas, and bomb differences kind of violation. Go to dashboard and chat channels. The audit store captures everything for legal traceability. This isn't a science project. These components are all off the shelf here. How a key out violation plays out. MA publishes an m, a update, rib added. Stream processor compares geometry hash key out strike rule. Engine flax, out of bound ECA and mca. Both get a slapping with a 3D overlay minute er, the change not base. Yes, that's the magic instant context rich feedback. Let's look at the numbers from pilot board changes. Revision lead time dropped from seven days to three days. Second example, part number approvals from four down to a day and half, and board spins fell from three to two. Those are scheduled wins. Your CFO can bank on the business side. Time to decision, snapshot count and conflict resolution all improve by 60 plus percent. Less churn means fewer Friday night scrambles. Lower expense and happier engineers finally, which were HR metrics will confirm, right? Observability can't be blind to itself. Grafana tracks events, volume and broker lag. Active workspace shows 3D Deltas. Kibana lets you grab every payload server, monitor config watches for DB connection losses and deadlocks ping ops only when threshold or breached. That's how we keep the pipeline healthy long after launch day. Roll out in four phases. Phase one, capture bomb outline events, low noise, instant value. Phase two, add constraints and variant events. Phase three, inline DFM, and thermal checks. Phase four feed. Suppliers and manufacturing within the same streams. Each phase is incremental. You harvest our ROI return investment. Before moving on. If your scars to save you, your paying, start small. Too many events too soon. Erodes trust. Govern schema versions. Yep. Nothing breaks. Integration faster than silent field changes. Align ECA and MCA naming conventions on day one. Invest early in broker observability like lag metrics or your early warning radar. Thank you. That's the journey from blind Spots to real time clarity. Thanks for writing up the room with me. Have a wonderful rest of the day.