Show HN: ScopeDB Cloud, Serverless event database with pipelined ScopeQL

Most teams start by writing business events and operational logs straight into their primary database or a basic log tool. It is the simplest way to get moving: there is no extra pipeline to run, and you can query whatever fields your application just wrote.

The trouble starts once the product gains traction. As request volume grows and features evolve, event tables swell and payloads change with every release. Answering everyday questions like which release introduced an error spike, where a workflow stalled, or where users drop off in a funnel begins to slow down your transactional database. Moving that history to a search engine quickly gets expensive to retain, while adopting a traditional analytical warehouse means locking down schemas upfront and maintaining a separate ETL pipeline.

ScopeDB Cloud is built so you do not have to give up day-one simplicity as your data grows. Applications append events with both fixed columns and evolving nested payloads directly into a single table, without a separate ETL pipeline. Durable data lives in shared object storage while isolated, stateless compute scales with your writes and queries, so you can move straight from high-level trends to the individual records behind them.

We see this firsthand with teams like Raft.build, where humans and AI agents build and debug software together and every room conversation sets off a stream of API requests, message routing, and runtime events. As usage grew, the team needed a place outside its transactional PostgreSQL database to retain high-concurrency traces and product events without adding an ETL pipeline. Today, Raft’s engineers and observability agents query ScopeDB as their shared source of evidence to verify releases and isolate failing operations.

Database errors spiked after the v1.21.23 rollout. Which queries are failing, and is it isolated to the new release?

Agent

messages.insert errors spiked on v1.21.23 (142 vs. 2 on v1.21.22).I'll count completed server.db.query spans over the last hour by version, database system, query name, and outcome, deduplicating each span once.

FROM raft.trace_events_v2 WHERE deployment_environment = 'production' AND span_name = 'server.db.query' AND row_kind = 'span_fact' AND event_time >= NOW() - 'PT1h'::interval AND event_time < NOW() DISTINCT ON service_name, trace_id, span_id GROUP BY service_version, db_system, query_name, outcome AGGREGATE count() AS operations ORDER BY operations DESC

Running…

Read the customer story: How Raft.build scales agent log analytics with ScopeDB

You do not need a complex outage to get value from your event data. Once you connect your agent to a ScopeDB workspace, it can inspect your tables, run read-only ScopeQL queries, and answer everyday questions about how your application is behaving:

Paste this prompt into Claude Code, Codex, or any agent with web and terminal access to install the ScopeDB skill and try these questions against sample data or your own events:

Install the ScopeDB skill from https://github.com/scopedb/scopedb-skills and help me get started.

Get started

ScopeDB Cloud is available for everyone today, so create your account, follow the quickstart, or start with your agent to load your first events and run a query in minutes. ScopeDB Cloud is currently free to use during public preview.