Find the slow request, then the slow span.

Q: What is application performance monitoring (APM)?

APM measures how fast your application responds and where the time goes. It is usually built on distributed traces: a record of each request as it moves through your services, split into timed spans.

Q: Is Squasher a full APM suite?

Not yet. Squasher has distributed traces, span waterfalls, p95 latency trends, a service map, release health, latency anomaly detection, real-user Web Vitals, and host metrics. It has no profiler, no database query view, and no threshold alerts on span duration. If you need those today, keep your APM and send the same traces to Squasher.

Q: How do I send traces to Squasher?

Export OTLP over HTTP to https://ingest.squasher.ai/v1/traces with your API key in the x-squasher-key header. JSON and protobuf both work, with or without gzip. gRPC does not. On Vercel, the Squasher integration adds a trace drain for you.

Q: Does Squasher instrument my code automatically?

Use the OpenTelemetry instrumentation for your language, and Squasher reads the spans it produces. The Squasher SDKs create spans in two places: Cloudflare Workers handlers wrapped with withSquasher, and Node.js handlers wrapped with withHttpRequest.

Q: How do traces connect to errors and logs?

A span that ends with an ERROR status is also stored as an error log with the same trace and span IDs. A trace lists its related logs, and a log with a trace ID opens its waterfall.

Q: Can I alert on latency?

Yes for Web Vitals, with a query monitor on p75, p95, or the share of poor samples, and for uptime checks that respond slowly. Not yet for span or trace duration.

Q: How long does Squasher keep traces?

Traces are kept 7 to 30 days, depending on your plan. They count toward your plan's telemetry volume together with logs and metrics.

Q: Can I sample traces?

The Squasher SDKs sample by trace ID, with separate rates for successful, failed, and slow requests. For tail sampling, run the OpenTelemetry Collector before you export.

Send OpenTelemetry traces and browser Web Vitals. Squasher charts p95 latency, draws each trace as a span waterfall, and links spans to logs by trace ID.

Start free

Free plan. No credit card required.

app.squasher.ai/acme/checkout-api/traces

POST/api/checkout8a71f0c24e9d4b1f9a3c6e2d7b5f0a14

Service: checkout-api
Duration: 2.31 s
Spans: 5
Status: Error

Span0 ms578 ms1.16 s1.73 s2.31 s

checkout-apiPOST/api/checkout

2.31 s

checkout-apicart.load

162 ms

pricingGET/prices

1.90 s

pricingdb.query prices

1.76 s

checkout-apiparseCurrency

4 ms

Example trace

Works with the tracing you already run.

OpenTelemetry
Node.js
Python
Go
Vercel

What you can see today.

Traces, latency, and real-user scores in the same project as your errors and logs.

Trace explorer
Filter by service, root span, duration, HTTP method, status code, or errors. Sort by duration to find the slowest.
Span waterfall
Search spans, show only errors, zoom up to 4×, and read each span's attributes, events, and links.
Latency trends
p95 trace latency, error rate, and request volume for the project, on the dashboard home page.
Web Vitals
LCP, INP, CLS, FCP, and TTFB from real users, from p50 to p99, by page, country, browser, and device.
Logs by trace ID
Each trace lists its related logs. Each log with a trace ID opens its waterfall.
Uptime response times
HTTP monitors record response time per check and region, and can mark a slow check degraded or failed.

Latency as a trend, not a guess.

The dashboard home page charts the numbers you check first when something feels slow.

p95 trace latency, error rate, and request volume, per time bucket.
Error rate is the share of traces with at least one error span.
squasher traces operations lists p50, p95, and error rate per operation.
Query spans with SQL from the CLI or MCP.

Traces API

app.squasher.ai/acme/checkout-api

Dashboard

Live traces, logs, metrics, and error health for this project.

Trace latency (p95)95th percentile trace latency per bucket.2.31 sLatest p95

Error rateShare of traces with at least one error span.4.8%Window rate

Request volumeTrace requests per bucket.48,210Last 1h

Example data

A failed span leads to its logs.

Slow and failed requests share one trace ID with the logs they wrote.

A span with ERROR status is also stored as an error log, with the same IDs.
Related logs on a trace lists every log line with its trace ID.
A log's Trace tab draws the waterfall around it.
With GitHub linked, spans with code attributes open the source file.

Error monitoring

Waterfalltrace 8a71f0c2

POST /api/checkout
cart.load
pricing.load
db.query prices
render.total
0 msrender.total threw2.31 s

Related logssame trace ID

14:06:09infoPOST /api/checkout 200 188ms
14:06:11infocart.load cart=8f21 items=3
14:06:12warnpricing: price pending at render
14:06:12errorTypeError: reading 'amount' of undefined
14:06:14infoPOST /api/checkout 500 2.31s
14:06:15warnretry pricing.load attempt=2
14:06:16infoGET /api/cart 200 84ms
14:06:18errorTypeError: reading 'amount' of undefined
14:06:09infoPOST /api/checkout 200 188ms
14:06:11infocart.load cart=8f21 items=3
14:06:12warnpricing: price pending at render
14:06:12errorTypeError: reading 'amount' of undefined
14:06:14infoPOST /api/checkout 500 2.31s
14:06:15warnretry pricing.load attempt=2
14:06:16infoGET /api/cart 200 84ms
14:06:18errorTypeError: reading 'amount' of undefined

Example trace and logs.

Real-user Web Vitals, on by default.

The browser SDK measures Core Web Vitals from every visit, with no extra code.

LCP, INP, CLS, FCP, and TTFB, rated with the standard thresholds.
Pick p50, p75, p90, p95, or p99. p75 is the default.
Break scores down by page, country, browser, or device.
Alert with a query monitor on p75, p95, or the share of poor samples.

Browser SDK

app.squasher.ai/acme/checkout-web/web-vitals

Web vitals

Find slow, unstable, and unresponsive experiences in real-user data.

Web Vital scoresP75 scores and quality distribution for this range.18,251 measurements · 24h

LCPLargest renderNeeds improvement

3.1 sP75

18,204 samples · 58% good

INPResponsivenessGood

164 msP75

11,930 samples · 81% good

CLSVisual stabilityGood

0.04P75

18,187 samples · 92% good

FCPInitial renderGood

1.4 sP75

18,240 samples · 79% good

TTFBServer responseNeeds improvement

920 msP75

18,251 samples · 64% good

Example data

Send your first trace.

Keep your OpenTelemetry setup and point it at Squasher. Add the browser SDK for Web Vitals.

Point your exporter at Squasher

OTLP over HTTP, as protobuf or JSON. gRPC is not supported, so set the protocol.

OTEL_EXPORTER_OTLP_ENDPOINT="https://ingest.squasher.ai"OTEL_EXPORTER_OTLP_HEADERS="x-squasher-key=$SQUASHER_API_KEY"OTEL_EXPORTER_OTLP_PROTOCOL="http/protobuf"

OpenTelemetry guide

Add the browser SDK
Initialize it once. Web Vitals, browser errors, and session replay start on their own.
```
npm install @squasher-ai/browser
```
Browser SDK
Find the slowest trace
Open Explore, switch to Traces, and sort by duration.

Paste into your coding agent

Install Squasher for this project. If Squasher skills or MCP are not configured in this agent environment, run `bash <(curl -fsSL https://squasher.ai/install.sh) --agents -y`. Then read https://skills.squasher.ai/squasher-onboard/SKILL.md, map every runnable app/service/worker/frontend/agent in the repo, choose the matching Squasher SDK, log drain, OTLP, or Agent SDK path, initialize with SQUASHER_API_KEY and SQUASHER_PROJECT_ID from environment variables, send one real verification event, and report the files changed plus the verification step. Do not hardcode or reveal real API keys.

bash <(curl -fsSL https://squasher.ai/install.sh) --agents -y

What is built, and what is not yet.

Squasher is not a full APM suite yet. Here is the line, so you can decide what to keep.

Distributed tracing

Status: Yes
Details: OTLP over HTTP, as protobuf or JSON. No gRPC.

Span waterfall

Status: Yes
Details: Search, errors only, zoom, attributes, events, and links.

Latency trends

Status: Yes
Details: p95 per trace on Home. p50 and p95 per operation in the CLI and API.

Real-user monitoring

Status: Yes
Details: Web Vitals and session replay.

Host metrics

Status: Yes
Details: CPU, memory, disk, and network from OpenTelemetry host metrics.

Service map

Status: Yes
Details: Services, their calls, and their databases, from your traces.

Latency alerts

Status: Partly
Details: Service p95 anomalies, Web Vitals, and uptime response time. No span duration threshold.

Browser-to-backend traces

Status: No
Details: The browser SDK does not send traceparent headers.

Profiling

Status: No
Details: Not available.

Database query insights

Status: No
Details: Not in the dashboard.

Which APM capabilities Squasher has today, which are only in the CLI and API, and which are missing.
Capability	Status	Details
Distributed tracing	Yes	OTLP over HTTP, as protobuf or JSON. No gRPC.
Span waterfall	Yes	Search, errors only, zoom, attributes, events, and links.
Latency trends	Yes	p95 per trace on Home. p50 and p95 per operation in the CLI and API.
Real-user monitoring	Yes	Web Vitals and session replay.
Host metrics	Yes	CPU, memory, disk, and network from OpenTelemetry host metrics.
Service map	Yes	Services, their calls, and their databases, from your traces.
Latency alerts	Partly	Service p95 anomalies, Web Vitals, and uptime response time. No span duration threshold.
Browser-to-backend traces	No	The browser SDK does not send traceparent headers.
Profiling	No	Not available.
Database query insights	No	Not in the dashboard.

Checked against the product on September 24, 2026.

Questions, answered.

What is application performance monitoring (APM)?

APM measures how fast your application responds and where the time goes. It is usually built on distributed traces: a record of each request as it moves through your services, split into timed spans.

Is Squasher a full APM suite?

Not yet. Squasher has distributed traces, span waterfalls, p95 latency trends, a service map, release health, latency anomaly detection, real-user Web Vitals, and host metrics. It has no profiler, no database query view, and no threshold alerts on span duration. If you need those today, keep your APM and send the same traces to Squasher.

How do I send traces to Squasher?

Export OTLP over HTTP to https://ingest.squasher.ai/v1/traces with your API key in the x-squasher-key header. JSON and protobuf both work, with or without gzip. gRPC does not. On Vercel, the Squasher integration adds a trace drain for you.

Does Squasher instrument my code automatically?

Use the OpenTelemetry instrumentation for your language, and Squasher reads the spans it produces. The Squasher SDKs create spans in two places: Cloudflare Workers handlers wrapped with withSquasher, and Node.js handlers wrapped with withHttpRequest.

How do traces connect to errors and logs?

A span that ends with an ERROR status is also stored as an error log with the same trace and span IDs. A trace lists its related logs, and a log with a trace ID opens its waterfall.

Can I alert on latency?

Yes for Web Vitals, with a query monitor on p75, p95, or the share of poor samples, and for uptime checks that respond slowly. Not yet for span or trace duration.

How long does Squasher keep traces?

Traces are kept 7 to 30 days, depending on your plan. They count toward your plan's telemetry volume together with logs and metrics.

Can I sample traces?

The Squasher SDKs sample by trace ID, with separate rates for successful, failed, and slow requests. For tail sampling, run the OpenTelemetry Collector before you export.

On squasher.ai

In the docs

Priced by volume, never by seat.

Start free. Invite the whole team. AI triage is included on every plan.

Compare every plan

Free

For side projects, evals, and internal tools.

$0/ month

No card needed. No overage charges.

10 GB telemetry
5K replays
500 AI credits

Included

AI triage on new issues
500 AI credits / month (an investigation uses 50)
1 project

Pro

For small production apps that want one tool, not five.

$29/ month

Flat price. No per-seat fees.

100 GB telemetry
10K replays
3K AI credits

Everything in Free, plus

3,000 AI credits / month (an investigation uses 50)
5 projects
30 monitors with 30-second checks

Team

Scale

For high-volume teams with security, procurement, and support needs.

$299/ month

Flat price. No per-seat fees.

1 TB telemetry
150K replays
30K AI credits

Everything in Team, plus

30,000 AI credits / month (an investigation uses 50)
100 status pages
250 monitors with 30-second checks

Need more volume, a security review, or custom terms? Get in touch.

Go squash some bugs.

Connect one service and send one real event. The crew takes it from there.

Start free

Free plan. No credit card required.

Find the slow request, then the slow span.

Works with the tracing you already run.

What you can see today.

Trace explorer

Span waterfall

Latency trends

Web Vitals

Logs by trace ID

Uptime response times

Latency as a trend, not a guess.

A failed span leads to its logs.

Real-user Web Vitals, on by default.

Send your first trace.

Point your exporter at Squasher

Add the browser SDK

Find the slowest trace

What is built, and what is not yet.

Distributed tracing

Span waterfall

Latency trends

Real-user monitoring

Host metrics

Service map

Latency alerts

Browser-to-backend traces

Profiling

Database query insights

Questions, answered.

What is application performance monitoring (APM)?

Is Squasher a full APM suite?

How do I send traces to Squasher?

Does Squasher instrument my code automatically?

How do traces connect to errors and logs?

Can I alert on latency?

How long does Squasher keep traces?

Can I sample traces?

Keep exploring.

On squasher.ai

In the docs

Priced by volume, never by seat.

Free

Pro

Team

Scale

Go squash some bugs.