Announcing Sazabi 1.0 and General Availability

Today, we're proud to announce the release of Sazabi version 1.0 and general availability. We're also launching two new features: Dashboards and Log Search. Lastly, we're excited to share that Sazabi is now self-serve. You can get started in less than ten minutes without talking to a human. Sign up at www.sazabi.com.
What is Sazabi?
Sazabi is the AI-native observability platform for fast-moving engineering teams.
In today's crowded dev tools market, it's perhaps easiest to understand Sazabi by explaining what it is not:
- It's not an AI SRE that sits on top of other tools and triages alerts.
- It's not an agent/LLM observability platform that allows you to inspect agent trajectories and build evals.
- It's not a coding agent designed to be used as your “daily driver” (like Claude, Codex, or Cursor).
Instead, Sazabi is a comprehensive solution for observability that supports any workload type - frontend, backend, database, infrastructure, or agent - and satisfies any use case, whether you're trying to gauge overall system health, investigate a specific failed request, or debug a rogue agent.
What does it mean that Sazabi is “AI-native”? Two things:
- AI at the core: Sazabi's main features are only possible because of AI. For example, it uses AI to automatically detect issues from your log stream, create intelligent indices on top of your data, and uncover blind spots in your instrumentation.
- Built for agents: Sazabi exposes programmatic interfaces for every operation, from sign-up to onboarding to connecting integrations to managing your subscription. To our knowledge, Sazabi is the first observability platform that can be used in a completely headless manner.
If you're a developer wondering why you should care about Sazabi, here's your answer: Sazabi solves observability for you, your entire stack, and your entire team in ten minutes or less.
With Sazabi, you don't have to worry about instrumentation, creating monitors, or maintaining dashboards. Just connect your log streams, and everything works. After setup, you can rest easy knowing you'll get a notification when something breaks in production. You and your team can get back to building.
Don't just take our word for it. Ask Abhi, CTO of Mastra, about how Sazabi is a “trusted safety net” for their team and saves every one of his engineers an hour per day. Or ask Andrew, CTO of Weave, how their team has never needed to adopt a “traditional” observability tool because they started with Sazabi.
Mastra and Weave are two of the incredible design partners we've had the privilege of working with over the last few months. The full case studies are available here.
Product Update
1.0 is first and foremost a product milestone.
The first update we'd like to share is that Sazabi's core feature, Automatic Issue Detection, is now ready for general use.
Here's how Automatic Issue Detection works:
- You send us logs.
- We analyze them, investigate anomalies, and send back rich, actionable alerts about real problems in your app.
Every Sazabi “issue” comes with a detailed description of the impact and root cause, as well as a suggested fix. We make it easy to hand this context to your local coding agent by copying a prompt or kick off a third-party cloud agent through one of our integrations. Importantly, Sazabi doesn't focus on code-generation. We aim to be the best agent for detection and root-cause analysis - not the best agent for coding.
We've spent months refining Automatic Issue Detection so that it is fast, accurate, and cost-effective. We're excited to see what you think!
At Sazabi, we believe detection and root-cause analysis are the heart of observability. But we also acknowledge that teams rely on observability tools for a range of use cases today, including some that Sazabi doesn't support… yet. That's why we're introducing two new features to address these gaps: Dashboards and Log Search.
Log-Based Dashboards
Everyone loves a good dashboard. They may not be the most useful for detecting or investigating issues, but they're exceptionally good at illustrating them. With Sazabi 1.0, you can now create beautiful live-updating Dashboards in the Sazabi web app.
Here's the kicker: you can use your logs. Instead of requiring you to instrument metrics separately or know the difference between a counter, a gauge, and a rate, Sazabi allows you to build beautiful, live-updating Dashboards on top of the log data you already send us.
The Fastest Log Search on the Planet
From the beginning, our team believed that Sazabi should be agent-agnostic: engineers should be able to use Sazabi with any coding agent they prefer. But while Sazabi offered agent-friendly interfaces for interacting with issues, it had no such interfaces for querying logs. This made it hard to do ad hoc analysis or go deeper on an issue that Sazabi surfaced.
No longer. Sazabi 1.0 comes with powerful Log Search capabilities through CLI, API, and MCP.
Not only that, but Sazabi's Log Search is the fastest on the planet. We leverage intelligent indices built on top of your raw data to perform Log Searches faster than any other database or vendor today. The result: faster answers, faster investigations, higher uptime, and token savings.
“Fastest on the planet” is not a claim we make lightly. Sazabi's engineering team will be releasing a blog post in the coming days with benchmarks and an explanation of the techniques that make this possible.
With Dashboards, Log Search, and Automatic Issue Detection, Sazabi can now credibly claim to be a complete observability solution - a one-stop shop for serious engineering teams that maintain production systems.
Lessons from Alpha and Beta
Sazabi announced our closed alpha in March and our open beta in June. What have we learned over the past seven months of building? Three lessons stand out.
First, we learned that building great infrastructure takes time.
Great infrastructure is cheap, reliable, simple, and scalable. Sazabi's core infra is our log intake, storage, and query system. We spent months iterating on this, disqualifying no fewer than five different architectures, before arriving at an implementation that meets our cost, latency, and durability requirements.
Our hard work is paying off. Today, Sazabi ingests over 10 GB of logs per minute and over 500 TB per month across the five regions we support. Teams that choose Sazabi can have confidence that our platform will scale gracefully, never drop data, and never become unavailable during a critical moment.
Second, we learned that the biggest and most important challenge in observability is instrumentation.
Put simply, you can't investigate an error that doesn't appear in the logs or monitor a service you don't know about. We've done a ton of work in-product to mitigate this:
- We have close to 200 integrations, including over 50 native integrations with leading cloud providers, making it extremely convenient to send us data.
- We scope-down the instrumentation problem by requiring users to send only one type of data: logs.
- We're compatible with both OpenTelemetry and proprietary log shippers like the Sentry SDK, PostHog SDK, and Datadog agent.
- We lean heavily into “agentic instrumentation”.
Nevertheless, instrumentation is still a headache for engineering teams and an opportunity for Sazabi to provide value. We're excited to keep working on this problem.
Lastly, we learned that Sazabi's killer feature and biggest differentiator is Automatic Issue Detection.
Anyone can point Claude Code at a Datadog alert and end up with the root cause. But that assumes someone took the time to set up this Datadog monitor to begin with. It also assumes your monitor coverage is comprehensive and up to date, which is unlikely given the pace of software development today. Meanwhile, it's slow and expensive to investigate each and every alert, especially when you consider false positives and duplicates.
Sazabi solves all of this by surfacing issues automatically without monitors, suppressing noise, and delivering every issue with all the context required for a fix. Our Automatic Issue Detection system is at least an order of magnitude cheaper than investigating alerts individually with Claude or Codex.
What's Next
1.0 is just the beginning. We've got so much more work to do.
In the coming months, the Sazabi team will be polishing and improving our existing core primitives, as well as investing more in agentic instrumentation so that future users can onboard with zero effort. We'll also be building deeper integrations with third-party coding agents to make hand-offs even easier.
Sazabi's ultimate goal is to unlock “self-healing software”: a world where software heals and improves itself without human intervention and direction. We believe the key is understanding what's happening in production.
To everyone who has supported us thus far: thank you. We're excited to continue working with you and to share Sazabi 1.0 with the world.
Sherwood Callaway
CEO & Founder
sherwood@sazabi.ai

