Founder @parseablehq | Helping teams leverage their telemetry data for better outcomes | Try now at https://xitter.cf/t.co/NSdrrjzP7S
SF / BLR
Joined October 2009
- Tweets3.4K
- Following1.3K
- Followers1.1K
- Likes4.3K
Pinned Tweet
Parseable is a telemetry datalake on object storage. Logs, metrics, traces, and AI agent telemetry go into one system, stored as open Parquet files on object store. You query with SQL or natural language. You keep data for 12 months instead of 30 days. You pay 80x lesser.
Some context on why we built this.
Observability costs grow linearly with data volume, but value doesn't. Most teams delete telemetry after a month because their platform makes retention irrational. AI workloads are making this worse — agents generate telemetry at extreme volume and cardinality, and that data is scattered across separate tools that don't correlate with each other.
We spent the last year talking to engineering leaders about this. 75% cited cost as their primary observability concern. 84% of organizations plan to consolidate tools in 2026. The pattern was clear: teams are paying more each year, retaining less data, and getting slower queries.
@parseablehq fixes the economics by fixing the architecture. Columnar Parquet format compresses telemetry up to 90%. Compute and storage scale independently. No pre-indexing, no write amplification, no JVM overhead. The entire engine is written in Rust and runs on @TigrisData , a globally distributed S3-compatible store with zero egress fees.
Pro plan is $0.39/GB ingested. 12 months retention. All AI features included — natural language queries, forecasting, anomaly detection. Unlimited hosts, users, dashboards. 14-day free trial, no credit card.
Enterprise gets bring-your-own-bucket for unlimited retention, data residency, BYOC or on-prem deployment, and Iceberg support.
Self-hosted is still available. Single binary. Same engine.
If you're running AI workloads alongside traditional infrastructure and tired of stitching together three tools to understand one incident, try out Parseable today at telemetry.new/
Crazy 24 hours for security teams - eMudhra, TiDB, PHP, GitLab, Railway.
It just feels like the beginning, we have a long way to go as we adjust to the new paradigms of ensuring security and privacy.
Nitish Tiwari retweeted
My cofounder Somesh @MathproBro is hosting a session Friday at 3 PM on the math behind AI: graphs, GNNs, transformers and deceptively simple questions about how they work.
If you want a path to work in applied math or at an AI lab, pls do attend. Link below
1. what?
We discovered a sophisticated recording device that was hidden inside one of our offices. We used our models to understand the device and where it had come from, identifying that a specific employee had planted it a few days earlier. We referred the matter to the FBI, and we’re grateful to the agents and prosecutors who moved quickly to make an arrest. The FBI’s investigation is ongoing and we will continue to support their work.
At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was terminated.
We take protecting our technology, data, and people seriously, and as a matter of national importance. Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.
complicated setup while self hosting observability tooling is one of the biggest problems, yet least talked about. So we built Parseable as a single binary, that can be deployed in minutes on your laptop. The same simplicity then carries over when you deploy to prod, with zero external dependencies and an object store first design.
Nitish Tiwari retweeted
100M time-series/min? That's not just big data, that's a *statement*. Open observability datalakes like @ParseableHQ are the ONLY way solo devs can truly own their stack without vendor lock-in. Essential for scaling AI/data products without breaking the bank. #microSaaS #buildinp
Agents work best when grounded in real data. pb cli allows unencumbered access to the data via linux native tooling. This way the agent doesn't need to find its way around a UI while debugging something - agents are faster and accurate with CLI
sed aka stream editor allows bulk changes to unstructured text. This combined with pb offers a great way to understand what is happening while you debug
CLI tools are very helpful in slicing data points once you've identified the data you want to look at. We wrote pb to allow users to really take control of their workflows. See how to pipe pb output to awk for processing.
Self hosting observability system often means letting go of the high end functionality like anomaly detection, or sometimes stitching together different components that are fragile.
We've bundled an accurate, zero shot time series prediction model within Parseable - so you get anomaly detection built-in without sending data out.
Canvas style exploration of observability data offers a clean way to segregate context across your questions
Nitish Tiwari retweeted
btw we're hiring for...basically everything!
Engineers (so many eng roles), design, deved/devrel, events, sales, solutions engineering, support.
planetscale.com/careers
Nitish Tiwari retweeted
For most products you can build persistence entirely on object storage now. Conditional writes closed the gap. Ampbase runs their whole control plane on Tigris this way.
Nitish Tiwari retweeted
If you've wanted to write for The Consensus but you don't have the time, or it seemed daunting, I'm looking for folks to write short briefs for weekly roundups.
Come write with me!
theconsensus.dev/contribute.…
Nitish Tiwari retweeted
Continuing on the WAL-on-S3 discussion, it was good to see @trylle actually conducting performance and correctness tests. There are improvements to be made in Tigris but we are getting there.
New blog post! Write-Ahead Logs on S3: Compatibility Is Not Performance. We measured S3 Express against S3, then against Tigris and of course Cloudflare R2.
trylle.com/blog/wal-s3
