Assertion Supported AI assessment confidence: 95% certainty 4/5 debate potential 2/5

Go's garbage collection makes InfluxDB unsuitable for sub-millisecond trading

Paul Dix · Paul Dix, InfluxDB // Open-Source Time Series Database // Data Driven NYC (FirstMark Capital) · Apr 2, 2015 · at 20:20

Paul Dix, founder and CEO of InfluxDB, explains why InfluxDB's Go architecture is not designed for sub-millisecond critical-path trading algorithms.

0:00 / 0:09exact quote · 9.8s

▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark

“This isn't, this is written in Go, so, which is a garbage collected language, so worst case response time can be worse than what you would want in that kind of setup.”

quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →

The claim, plainly

Because InfluxDB is written in Go, a garbage-collected language, its worst-case response times make it unsuitable for sub-millisecond trading algorithms.

a quote alone can be ambiguous, so this restates it with the references and context of the conversation around it resolved. AI-written; the tape below the player is the source.

Assessment

The resolution criteria require confirmation of two facts: first, that InfluxDB's public source code confirms it is written in Go; second, that technical documentation or benchmarks confirm Go garbage collection pause times prevent guaranteed sub-millisecond worst-case latency bounds. InfluxDB's public repository (github.com/influxdata/influxdb) confirms it is implemented in Go. Technical documentation and runtime analyses confirm Go GC does not offer hard real-time guarantees and introduces non-deterministic tail latency spikes (e.g., via mark assist under load), preventing guaranteed sub-millisecond worst-case latency bounds required for sub-millisecond algorithmic trading.

InfluxDB official repository (github.com/influxdata/influxdb) confirms implementation in Go.
Go runtime documentation and benchmarks confirm lack of deterministic real-time latency bounds, with GC background work (mark assist) causing non-deterministic tail latency spikes above 1 millisecond.

Checked against

[1] go.dev [2] gopheracademy.com [3] dev.to [4] stackoverflow.com [5] support.tools [6] google.com [7] reddit.com [8] reddit.com

"sub-millisecond" "garbage collection" Go trading latency worst-case "Go 1.5" "garbage collection" pause times 10ms InfluxDB source code Go language GitHub Go GC latency worst-case guaranteed sub-millisecond hard real-time "Go 1.8" "GC pause" microsecond millisecond bounds Go GC pause times millisecond low latency trading benchmark Go garbage collection pause times worst case latency sub millisecond trading

assessed as of 2026-07-30

This assessment is an AI reading of a public statement, made against the cited references with the confidence shown. It can be wrong. The quote, the tape and the references are all on this page; if something is off, use report an error below. full disclaimer →

Mentioned

Go InfluxDB

⚑ Spot an error? Report it

More from Paul Dix

Prediction Not checkable as stated

Paul Dix: Nobody will ever make the Lisp programming language popular

“Lisp, we're not going to do because Paul Graham and Rich Hickey couldn't make Lisp popular, and neither will we. Nobody will.”

Paul Dix Jun 12, 2019 ▶ 22:16 What's Next for Open-Source Time Series Data? // Paul Dix, Influx Data (FirstMark's Data Driven NYC)

Insight

Why traditional databases fail at time series workloads

“And actually no Regular database is designed with this kind of workload in mind. Databases, for the most part, assume that you want to keep your data around for a very long time. So, these kind of unique aspects of time series make it kind of a degenerate case…”

Paul Dix Jun 12, 2019 ▶ 5:47 What's Next for Open-Source Time Series Data? // Paul Dix, Influx Data (FirstMark's Data Driven NYC)

Assertion Supported

Flux turns InfluxDB into a serverless execution platform for time series

“It essentially turns the database into a serverless execution platform for time series data. The idea is you can define any sort of custom logic that you want, inject it into the database, and it will periodically run that logic over the data that you're writi…”

Paul Dix Jun 12, 2019 ▶ 17:56 What's Next for Open-Source Time Series Data? // Paul Dix, Influx Data (FirstMark's Data Driven NYC)

Assertion Not checkable as stated

Paul Dix: Telegraf is likely on tens to hundreds of millions of servers

“This project actually is our most popular open source project by far. We don't have any sort of tracking on it right now, but I would estimate that it's probably deployed on tens of millions of servers across the world, if not hundreds of millions at this poin…”

Paul Dix Jun 12, 2019 ▶ 7:44 What's Next for Open-Source Time Series Data? // Paul Dix, Influx Data (FirstMark's Data Driven NYC)

Disclosure

InfluxData is building PromQL support directly into Flux

“And we also are in the middle of building in support for PromQL, the Prometheus query language.”

Paul Dix Jun 12, 2019 ▶ 15:46 What's Next for Open-Source Time Series Data? // Paul Dix, Influx Data (FirstMark's Data Driven NYC)

Insight

DevOps metrics and IoT sensor data spaces look surprisingly similar

“And those, the sensor data and the DevOps data spaces, they look surprisingly similar. Because when you think about DevOps data, and you think about the metrics that you're collecting, the sensors are just software that you have on your servers. Whereas in sen…”

Paul Dix Apr 2, 2015 ▶ 1:52 Paul Dix, InfluxDB // Open-Source Time Series Database // Data Driven NYC (FirstMark Capital)

Go's garbage collection makes InfluxDB unsuitable for sub-millisecond trading

More from Paul Dix

Turn any episode into a week of clips.