Hookdeck

𝗪𝗲 𝗺𝗮𝗱𝗲 𝗼𝘂𝗿 𝗳𝘂𝗹𝗹 𝗺𝗲𝘁𝗿𝗶𝗰𝘀 𝗱𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 𝗳𝗿𝗲𝗲. 𝗕𝗲𝗰𝗮𝘂𝘀𝗲 𝘁𝗿𝗲𝗻𝗱𝘀 𝘁𝗲𝗹𝗹 𝘆𝗼𝘂 𝘁𝗵𝗶𝗻𝗴𝘀 𝗶𝗻𝗱𝗶𝘃𝗶𝗱𝘂𝗮𝗹 𝗲𝘃𝗲𝗻𝘁𝘀 𝗻𝗲𝘃𝗲𝗿 𝘄𝗶𝗹𝗹. Real workspace data: 33,955 requests. 29,710 accepted. 4,245 rejected. 4,245 discarded. This isn't about debugging individual failures - it's about seeing patterns across your entire webhook lifecycle: → 𝗦𝗼𝘂𝗿𝗰𝗲𝘀: Are rejections climbing from a specific provider? (API key rotation?) → 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗶𝗼𝗻𝘀: Failure ratio spiking on one connection? (Specific endpoint degrading) → 𝗗𝗲𝘀𝘁𝗶𝗻𝗮𝘁𝗶𝗼𝗻𝘀: Events piling up? Latency increasing? (Your endpoint can't keep up with load) Those purple spikes at 20:00? That's your normal. Until one day it isn't. And that deviation tells you something changed upstream before your customers notice. The power isn't in seeing that 4,245 requests were rejected - it's in seeing WHEN they were rejected, from WHICH sources, through WHICH connections, to WHICH destinations, and how that changes over time. This was previously only available with our growth plan at $499/month. But we believe teams with better visibility make better decisions. When you can see patterns and trends across sources, connections, and destinations, you can optimize proactively instead of reacting to issues. Now free: - Full lifecycle metrics (sources → connections → destinations) - Multiple dimensions and filtering - Time-based patterns and trends - Real-time updates The best infrastructure decisions come from understanding actual behaviour, not assumptions. Most teams are surprised by what their real traffic patterns reveal. [link to announcement in the comments]

We just published a tutorial on building an outbound webhook starter with Hookdeck and Encore - with zero infrastructure setup. The tutorial covers: - Reliable webhook delivery with automatic retries - Webhook endpoint management and testing - Real-time delivery tracking and monitoring  - Automated infrastructure setup with Encore - React frontend for webhook configuration - Integration with Hookdeck's Event Gateway Hookdeck handles delivery reliability and webhook infrastructure complexity while Encore automatically takes care of the API, database, and deployment. Perfect for SaaS platforms that need to send webhook notifications to users when events occur - like deployments completing, payments processing, or user actions. Tutorial: https://lnkd.in/gC-xv6Yg

Hookdeck Metrics are now free - on every plan, including free. We believe strong observability should be a baseline, not a premium feature. So we’ve removed the paywall on Metrics and made a few UX improvements along the way. Now you’ll get: ✅ Full visibility into requests, events, and delivery attempts ✅ Integrated totals within each chart ✅ A clearer, more intuitive grouping layout ✅ Chart “Explore” option to quickly dive from metrics to raw request data

One year ago, we made a bet that PostgreSQL was holding us back. Today, that bet paid off with a 40% reduction in infrastructure costs. This marks the final piece of the biggest platform shift in Hookdeck's history: moving to eventual consistency, while maintaining a strongly consistent event store at the core. Everything else in our system is now eventually consistent. The key insight? We needed to move away from PostgreSQL for our events lifecycle. As much as we love Postgres, using it for high-throughput event processing at scale was like using a Swiss Army knife when you need a scalpel. Here's what we actually did: 𝗗𝗼𝘂𝗯𝗹𝗲𝗱 𝗱𝗼𝘄𝗻 𝗼𝗻 𝗖𝗹𝗶𝗰𝗸𝗛𝗼𝘂𝘀𝗲 - We were already using it, but now it's central to our event storage strategy. The columnar storage and compression ratios are perfect for append-heavy event data. 𝗞𝗮𝗳𝗸𝗮 𝗮𝗻𝗱 𝗗𝗿𝗮𝗴𝗼𝗻𝗳𝗹𝘆 𝗯𝗲𝗰𝗮𝗺𝗲 𝗼𝘂𝗿 𝗯𝗮𝗰𝗸𝗯𝗼𝗻𝗲𝘀 - We're leveraging Kafka significantly more for event streaming and decoupling services. We massively scaled up our Dragonfly usage. This was a game-changer. Dragonfly is an absolute beast, handling millions of operations per second without breaking a sweat. For hot-path data and caching layers in an eventually consistent system, it's unmatched. The real power of moving to eventual consistency? It forced us to ruthlessly evaluate what actually belongs in the critical path. By moving everything non-critical out of the request flow, we didn't just save money... We dramatically improved performance. This migration took a full year to design, implement, and roll out safely. Every step had to be carefully orchestrated to maintain our reliability guarantees while fundamentally changing how data flows through our system. Seeing this chart for the first time after the migration completed? Pure validation. That cliff drop isn't just a cost reduction. It's proof that we made the right architectural call at the right time. But here's what really matters: this directly improved our most critical metric: delivery p99. We measure this as the time from receiving a request to sending it back out (assuming no rate limits). 𝗢𝘂𝗿 𝗽𝟵𝟵 𝗶𝘀 𝗻𝗼𝘄 𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝘁𝗹𝘆 𝘂𝗻𝗱𝗲𝗿 𝟮 𝘀𝗲𝗰𝗼𝗻𝗱𝘀. We're working on making these performance metrics publicly available soon. The move to eventual consistency at scale isn't just about saving money. It's about building a platform that's both cost-efficient and blazing fast for our users. #DistributedSystems #SoftwareArchitecture #TechLeadership #StartupEngineering

July Hookdeck Review 🚢 - We launched the v4 Dashboard, redesigned for speed and clarity, with dedicated views for Sources, Destinations, and Transformations - You can now set the order of transformations and filters, giving you more control over processing and routing - Outpost v0.4.0 adds support for Azure Service Bus queues and topics - We partnered with Stripe Developer Relations to share best practices for rate-limit–friendly webhook patterns - We added new Source Types to the Event Gateway, including OpenAI, monday.com, and more

🚢 Outpost 0.4.0 released Outpost is our open source, self-hosted event delivery service. It gives you full control over how events are delivered as webhooks and directly to queues, with retries, delivery logs, and multi-tenant routing built in. In this release, we've added support for Azure Service Bus queues and topics as a destination type, ideal for teams building on Azure or developer tool companies supporting Microsoft-based stacks. 🔄 Highlights: - Azure Service Bus support (queues and topics) - New SDKs (TypeScript, Go, and Python, generated with Speakeasy) - Azure deployment scripts - Example showing how to create a tenant, add a destination, and publish an event with the SDKs

🆕 Hookdeck Event Gateway Feature: Control transformation rule order in your event flows. Choose exactly when your transformations run, before or after your filters. This unlocks more powerful and efficient event flows: ✅ Write focused transformations that only run when needed ✅ Filter using internal metadata, then clean it up before delivery ✅ Reduce noise and payload size downstream ✅ Align logic more closely to your app’s architecture You now control the rule execution order via both the dashboard UI and the API. Hookdeck runs your rules exactly in the order you define. Why does this matter? It gives you full control over when filtering and transformations happen, letting you design cleaner, more intentional workflows.

Webhooks are not a trivial problem at scale. When you're processing thousands of events per second, your strategy matters. Webhooks are usually delivered at-least-once. Assume duplicates and out-of-order events are the norm, not the exception. Idempotency isn't just a nice-to-have - it's critical. Your webhook handler needs to handle the same event multiple times without side effects. Synchronous webhook processing? Recipe for disaster. You'll hit timeouts, dropped connections, and potentially trigger retry storms. Event-driven architecture with robust queuing is your friend. Decouple, buffer, and process webhooks reliably.

Event-Driven Architecture: Beyond the Basics Choosing the right infrastructure for event-driven systems is a strategic move that can define your application's performance, scalability, and responsiveness. Key considerations include: - Selecting appropriate message brokers - Understanding distributed streaming platforms - Designing flexible event gateways - Creating resilient microservices architectures The goal? Build systems that react in real-time and adapt to changing business needs.

Our new dashboard v4 is a serious upgrade. It's a fundamental rebuild for clarity and speed. We tackled the core challenge of making event metrics instantly readable. Resource management used to feel cramped. Now? Your core workflows have room to breathe. We built this base so future iterations can be even faster and more intuitive. Data visualization isn't just about looking good. It's about giving you instant insight into your event streams. Every pixel is now purposefully designed to help you understand system behavior faster. Performance: We optimized loading times, streamlined navigation, and created a dashboard that feels responsive from the first click. We've reimagined how developers interact with complex event infrastructure. Clean, fast, informative.