How Depot merged 53 Polylane fixes in its first month
Depot builds software delivery infrastructure for engineers and AI agents, from source control and container builds to CI, caching, and registries. Thousands of engineering teams, including PostHog and PlanetScale, use Depot to build, test, and ship software faster.
Depot uses Polylane across AWS, PlanetScale and ClickHouse to detect issues, find the root cause and open the fix, so the team reviews pull requests instead of chasing alerts.
- cloud resources monitored in the first month
- 150.3K
- alerts triaged in the first month
- 1,074
- Polylane pull requests merged in the first month
- 53
- median from detection to a fix ready for review in the first month
- 24 min
cloud resources monitored in the first month
alerts triaged in the first month
Polylane pull requests merged in the first month
median from detection to a fix ready for review in the first month
Connected to Polylane
Cloud
AWS 
PlanetScale 
ClickHouse Tools
Honeycomb 
Axiom 
Sentry 
GitHub 
Linear 
Slack 
Devin 
Cursor 
MCP Server The challenge
Depot's customers notice problems through their own CI: a build that takes longer than it should, a runner that fails to start, a cache that stops hitting. By the time a customer reports it, their team has already lost time.
Depot's fleet is large. In its first month, Polylane mapped 150,301 cloud resources across AWS, PlanetScale and ClickHouse. Watching that by hand does not scale, and dashboards in Honeycomb and Axiom tell you something changed, not why.
“Our customers' builds run on our infrastructure, so when something drifts it shows up in their CI before it shows up anywhere else. Polylane watches the whole fleet, finds the cause and opens the fix. Most mornings the work is already done and waiting for review.”
Jacob Gillespie Co-founder & CTO at Depot Why Polylane
Polylane connects to Depot's cloud accounts, repositories and telemetry, and learns what normal looks like for each part of the fleet. When something drifts, it raises an issue with the evidence, investigates across code, deploys and infrastructure, and opens a pull request with the fix.
Depot merged its first Polylane fix nine days after connecting.
How Depot uses it
Polylane runs alongside Depot's existing observability in Honeycomb, Axiom and Sentry. Issues land in Slack and Linear with the root cause attached, and fixes land in GitHub ready for review. Every pull request, from engineers and from coding agents like Devin and Cursor, also gets a Polylane review against live production: 1,074 in the first month.
Depot's coding agents read Polylane's context over MCP, so the agent writing a change knows what that change touches in production.
The results
In the first month, Polylane monitored 150,301 cloud resources, detected 752 anomalies and surfaced 217 issues worth investigating. It opened 166 fixes with a median of 24 minutes from detection to a pull request ready for review, and Depot merged 53 of them.