Overview
Vercel’s drains forward observability data to an external destination in real time. Routing this data into Bronto lets you correlate frontend performance, user behaviour, and backend telemetry in a single place, with long-term retention and unified alerting across your full stack.
| Drain | What it contains | Schema reference |
|---|
| Logs | Runtime logs from your Vercel deployments | Log Drains reference |
| Traces | Distributed trace spans across your serverless functions | Traces reference |
| Speed Insights | Core Web Vitals and real user performance metrics | Speed Insights reference |
| Web Analytics | Visitor events and page view data | Analytics reference |
Prerequisites
- A Vercel Pro or Enterprise plan. See Vercel’s drain documentation for plan details.
- A Bronto API key with at least the Ingestion role. See API Keys for instructions on creating one.
Drain Types
Vercel exposes four drain types. Each maps to a separate Bronto ingestion endpoint and encoding format:
| Drain | Endpoint URL | Encoding |
|---|
| Logs | https://ingestion.<region>.bronto.io/ | NDJSON |
| Traces | https://ingestion.<region>.bronto.io/v1/traces | Protobuf |
| Speed Insights | https://ingestion.<region>.bronto.io/ | NDJSON |
| Web Analytics | https://ingestion.<region>.bronto.io/ | NDJSON |
<region> with your Bronto account’s region — either eu or us.
All four drain types use the same set of required headers:
| Header | Value |
|---|
x-bronto-api-key | Your Bronto API key |
x-bronto-dataset | The dataset name to write into |
x-bronto-collection | The collection name to write into |
The steps below apply to all four drain types. Repeat this process for each type you want to enable.
- Navigate to your team settings in Vercel and go to the Drains section
- Click Add Drain in the top-right corner
- Select the drain type you want to configure (Logs, Traces, Speed Insights, or Web Analytics)
- Configure the drain name and select which projects should be associated with it
All data associated with a single drain configuration will be grouped in the same dataset. We recommend creating separate drains for different projects and environments.
- For Logs drains: select the log sources and environments (e.g., production, preview) you want to include
- Most customers send all data to Bronto without sampling — Bronto is designed to handle high-volume ingestion. Sampling rules are available if you need to control Vercel drain costs.
- Select Custom Endpoint and set the URL to the appropriate endpoint from the table above
- Select the encoding format from the table above (NDJSON or Protobuf)
- Add the three required headers (
x-bronto-api-key, x-bronto-dataset, x-bronto-collection)
- Click Test — if everything is configured correctly you should see a 200 status response
If you receive an error, double-check that the endpoint URL is correct for the drain type, the API key has the Ingestion role assigned, and all three required headers are present and correctly formatted. If issues persist, contact Bronto Support.
- Click Create Drain to finish setup
Verify Data Flow
Once a drain is created, data from the selected projects and environments will begin flowing into Bronto. Navigate to your dataset in Bronto to confirm data is arriving as expected.
Out-of-the-Box Monitoring with Bronto Express
Once your Vercel drains are flowing into Bronto, you can use Bronto Express to instantly provision pre-built dashboards, monitors, and saved searches — no manual setup required.
Bronto Express detects which drain types you have configured and offers only the content that matches your data. If you later add additional drains, re-run Express provisioning to unlock the corresponding dashboards.
Bronto Express for Vercel is currently in preview and available on request. Please reach out to support to enable it for your account. Each of the three Vercel dashboards maps to a different drain type. You can ship with any subset — a Vercel Express package is considered valid if at least one drain’s content is fully provisioned.
Dashboards
Vercel Overview (Logs drain)
A high-level view of request traffic, error rates, cache behaviour, and firewall activity across your Vercel deployments.
| Widget | Type | What it shows |
|---|
| Total Requests | Query Value | Count of all proxied requests |
| 5xx Error Count | Query Value | Count of server error responses |
| Lambda Crashes | Query Value | Fatal-level events and status code –1 |
| Requests Over Time | Line | Proxied request volume over time |
| Requests by Source | Bar | Breakdown by source (build / edge / lambda / static / …) |
| HTTP Status Breakdown | Bar | Request count grouped by HTTP status code |
| Requests by Method | Bar | Request count grouped by HTTP method |
| Cache State Distribution | Bar | Breakdown by proxy.vercelCache (HIT / MISS / STALE / …) |
| Top Error Paths | Top List | Paths with the most 5xx responses |
| Top Edge Regions | Top List | Most active executionRegion values |
| Firewall Actions | Bar | Request count grouped by WAF action |
| Top Denied Clients | Top List | Client IPs most frequently denied by the firewall |
| Build Failures Over Time | Line | Build-source error log volume over time |
Vercel Web Vitals (Speed Insights drain)
Core Web Vitals at P75, segmented by route and device type, using Google’s published “good” thresholds.
| Widget | Type | What it shows |
|---|
| LCP P75 | Query Value | Largest Contentful Paint at the 75th percentile |
| INP P75 | Query Value | Interaction to Next Paint at the 75th percentile |
| CLS P75 | Query Value | Cumulative Layout Shift at the 75th percentile |
| TTFB P75 | Query Value | Time to First Byte at the 75th percentile |
| Web Vitals Over Time (P75) | Line | P75 trend for all metric types |
| LCP P75 by Route | Top List | Slowest routes by LCP P75 |
| INP P75 by Route | Top List | Slowest routes by INP P75 |
| Device Type Breakdown | Bar | Measurement count by device type |
Vercel Web Analytics (Web Analytics drain)
Visitor and engagement data from your Vercel Web Analytics drain.
| Widget | Type | What it shows |
|---|
| Pageviews Over Time | Line | Pageview event volume over time |
| Top Pages | Top List | Most visited paths |
| Top Custom Events | Top List | Most frequent custom event names |
| Referrer Breakdown | Bar | Pageviews grouped by referrer |
| Country Distribution | Tree Map | Pageviews grouped by country |
| Device Type | Bar | Pageviews grouped by device type |
Monitors
Monitors run continuously against your incoming data and fire alerts when thresholds are crossed. All windows are 5 minutes.
| Monitor | Drain | Severity | Threshold |
|---|
| Lambda Error Spike | Logs | Critical | 10 events |
| Lambda Crash | Logs | Critical | Any event with status code –1 |
| 5xx Error Rate Spike | Logs | Critical | 5% of total requests |
| High 4xx Rate | Logs | Warning | 15% of total requests |
| Firewall Deny Spike | Logs | Warning | 100 deny events |
| LCP P75 Regression | Speed Insights | Warning | P75 > 2,500 ms |
| INP P75 Regression | Speed Insights | Warning | P75 > 200 ms |
Thresholds for LCP and INP match Google’s and Vercel’s published “needs improvement” cut-off. You can adjust these in the monitor settings to match your own performance baseline.
Saved Searches
Pre-built filters you can open directly from the search bar or use as a starting point for your own investigations.
| Name | Drain | What it finds |
|---|
| Lambda Errors | Logs | Error and fatal log events from lambda sources |
| Lambda Crashes | Logs | Events with status code –1 (process crash) |
| Firewall Denies | Logs | Requests denied by the Vercel WAF |
| Build Failures | Logs | Error-level events from the build source |
| Poor LCP Pages | Speed Insights | LCP measurements above 2,500 ms |
| Poor INP Pages | Speed Insights | INP measurements above 200 ms |
