Migrate from Sentry

Squasher accepts the data that Sentry SDKs send, so you can move to Squasher by changing the DSN. Errors, traces, logs, metrics, sessions, and cron check-ins all arrive. You can also connect a Sentry organization, so that Sentry issues open incidents in Squasher while you compare the two tools.

What moves and what does not

In Sentry	In Squasher	Notes
Error events (`captureException`)	Issues	Squasher keeps the stack trace, error type, level, release, environment, user, and request.
Messages (`captureMessage`)	Log events	Only `error` and `fatal` events open issues.
Tags, contexts, extra data, breadcrumbs	Event attributes	Squasher keeps them as JSON on the event. Tags are also `sentry.tag.<name>` attributes. The dashboard does not show breadcrumbs.
Transactions and spans	Traces	Span `op`, `description`, `data`, and status become span names, kinds, attributes, and OpenTelemetry statuses.
AI spans (`gen_ai.*`)	AI observability	Squasher reads the model and token counts, and computes the cost.
Logs (`Sentry.logger`)	Logs	Logs do not open issues. The SDK sends each error as an error event, which opens the issue.
Metrics (`Sentry.metrics`)	Metrics	Counters are sums, gauges are gauges, and distributions are histograms.
Sessions (release health)	The `sentry.sessions` metric	Each session counts once as `started` and once as `exited`, `errored`, `crashed`, or `abnormal` (`sentry.session.status`).
Web vitals of browser page loads	Web Vitals	LCP, FCP, CLS, INP, and TTFB, with the page URL.
Cron check-ins (`captureCheckIn`)	Log events	One log per check-in. A heartbeat monitor named like the monitor slug gets each `ok` and `error`.
Issue grouping and custom fingerprints	Squasher grouping	Squasher groups events with its own rules. Sentry fingerprints do not carry over.
Source maps	Applied when read	Upload them with `sentry-cli`, a Sentry bundler plugin, or `squasher sourcemaps upload`. Squasher reads the debug IDs that the Sentry JavaScript SDK sends.
Attachments, replays, profiles, user feedback	Not stored	Squasher accepts them, so the SDK does not retry, and stores nothing.
Past issues and events	Not imported	Squasher starts with the events that you send after the switch.
Alert rules	Notification routes	Recreate them by hand. See Notifications.

Squasher stores Sentry data in the same tables, with the same OpenTelemetry attribute names, as data from OpenTelemetry and the Squasher SDKs. Queries, dashboards, and alerts work the same way for all of them. Squasher renames old attribute names, such as db.system and http.method, to the current OpenTelemetry names, and keeps the original.

Run both tools side by side

Keep Sentry running while you check Squasher. Use one or both of these:

Connect your Sentry organization. Each new Sentry issue opens an incident in Squasher, and Sentry keeps all its data. See Connect a Sentry organization.
Move one service at a time. Point one app or one environment at Squasher, and leave the rest on Sentry. Compare the two for a week that includes a deploy.

Point the Sentry SDK at Squasher

Set the DSN to this value in every Sentry SDK:

https://<ingest key>@ingest.squasher.ai/1

Use a Squasher ingest key (sq_pk_...), not a Sentry key. The ingest key selects your project, so the number at the end of the DSN can be any number. Sentry SDKs need a number there.

JavaScript
Python
Other SDKs

instrument.ts

import * as Sentry from "@sentry/node";

Sentry.init({
  dsn: `https://${process.env.SQUASHER_API_KEY}@ingest.squasher.ai/1`,
  environment: process.env.NODE_ENV,
  tracesSampleRate: 1.0,
  enableLogs: true,
});

The same options work in @sentry/browser, @sentry/nextjs, and the other JavaScript SDKs. In the browser, read the key from your public environment variables.

app.py

import os
import sentry_sdk

sentry_sdk.init(
    dsn=f"https://{os.environ['SQUASHER_API_KEY']}@ingest.squasher.ai/1",
    environment="production",
    traces_sample_rate=1.0,
    enable_logs=True,
)

Go, Java, .NET, PHP, Ruby, Elixir, and the mobile SDKs read the SENTRY_DSN environment variable when your code does not set a DSN:

export SENTRY_DSN="https://${SQUASHER_API_KEY}@ingest.squasher.ai/1"

If your code passes a DSN to init, change that value instead.

Then capture one exception and select Issues in the Squasher sidebar. Traces, logs, and metrics appear under Traces, Logs, and Metrics.

Squasher reads gzip and deflate compressed envelopes, so you can keep request compression on in the Sentry SDK. An envelope can be at most 1 MiB, before and after decompression. Brotli (br) and zstd envelopes return 415. Older SDKs that send to /api/<project>/store/ also work.

If you already set the tunnel option to a Squasher URL, you can keep it or remove it. With a tunnel, the SDK sends each envelope to the tunnel URL as it is. sentry-cli works with Squasher too. send-event and monitors run send to the DSN, and sourcemaps upload and the Sentry bundler plugins upload source maps to https://api.squasher.ai. See Upload with sentry-cli or a Sentry bundler plugin.

Issues and logs of the same error

Squasher opens issues only from Sentry error events, as Sentry does. A Sentry SDK also sends some errors as a log or as a failed span: for example, the Python logging integration sends logging.error() as an error event and as a log. Squasher stores the log and the span, but they do not open a second issue or count twice toward your error quota.

Replace the Sentry SDK

In a JavaScript app, you can remove the Sentry SDK and install a Squasher SDK instead: Next.js, Node.js, or Browser. The calls map like this:

Sentry	Squasher
`Sentry.init({ dsn })`	`init({ apiKey, projectId })`
`Sentry.captureException(error)`	`await captureError(error)`
`Sentry.captureMessage(message)`	`await captureMessage(message)`
`Sentry.setUser(user)`	`setUser(user)`. In `@squasher-ai/nextjs`, use `getClient().setUser(user)`.
`Sentry.setTag(key, value)`	`setTag(key, value)`. In `@squasher-ai/nextjs`, use `getClient().setTag(key, value)`.
`Sentry.addBreadcrumb(crumb)`	`addBreadcrumb(crumb)`. The SDK keeps breadcrumbs in memory, but Squasher does not store them.

Then delete the Sentry config files, such as sentry.server.config.ts and the withSentryConfig wrapper in next.config.js.

Connect a Sentry organization

When you connect a Sentry organization, Sentry sends issue updates to Squasher. A new Sentry issue opens an incident in your Squasher project, and the incident resolves when the Sentry issue resolves. Open in Sentry on the incident opens the issue in Sentry. Agents can also read Sentry issues and their latest events through the API. Each Squasher project connects to one Sentry project. To connect in the dashboard, open Integrations, select Connect on the Sentry card, and approve the installation in Sentry. If the organization has more than one project, choose it in Sentry project on the card when Squasher opens again. If the Sentry card shows Manual setup, organization connect is not available yet, and the API returns 503. Point the Sentry SDK at Squasher instead. To connect with the REST API instead, use a credential with the project:write permission, which an ingest key does not have. These examples read it from SQUASHER_TOKEN.

Create a connection link

curl -X POST "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/provider-connections/sentry/connect-plan" \
  -H "Authorization: Bearer $SQUASHER_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{}'

The response has the link in action.url. The link works for 10 minutes.

Install the app in Sentry

Open the link, choose your Sentry organization, and approve the installation. If the organization has one project, Squasher connects it and opens the Squasher dashboard.

Choose the Sentry project

If the organization has more than one project, Squasher opens the dashboard with a selection_token in the URL, and the Sentry card shows a project picker. To choose through the API instead, send the token and the slug of the Sentry project:

curl -X POST "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/providers/sentry/select-project" \
  -H "Authorization: Bearer $SQUASHER_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"selection_token": "<token from the URL>", "sentry_project_slug": "checkout-api"}'

Check the connection

curl "https://api.squasher.ai/v1/projects/$SQUASHER_PROJECT_ID/provider-connections/sentry" \
  -H "Authorization: Bearer $SQUASHER_TOKEN"

The connection works when status is connected.

To disconnect, send DELETE to the same URL. Squasher then ignores new Sentry events, but it cannot remove the app from Sentry. Uninstall the app in your Sentry organization settings too.

Check parity

Compare the service that you moved, over the same number of days before and after the switch:

The number of error events each day.
The top issues. The same errors must not be split into many issues or merged into one.
The release and environment values on each event.
The notifications. Each important error must reach the same channel as before.

Turn off Sentry

Recreate the Sentry alert rules as notification routes.
Point the source map step in your CI at Squasher, or replace it with squasher sourcemaps upload. See Source maps.
Create a heartbeat monitor for each Sentry cron monitor, with a name that gives the same slug (for example Nightly Sync for nightly-sync). Sentry check-ins then signal it, and it alerts when a job fails or does not check in.
If you connected a Sentry organization, disconnect it and uninstall the app in Sentry.
Remove the Sentry SDK, or keep it pointed at Squasher.

Give the migration to a coding agent

Move this app's Sentry reporting to Squasher. Keep the Sentry SDK and its options, including
tracing and logs. Set the DSN to https://<SQUASHER_API_KEY>@ingest.squasher.ai/1, with the key read
from an environment variable; the number 1 is required by the SDK and Squasher ignores it. Remove a
tunnel option that points at Sentry. Send one test error and confirm that it appears under Issues,
and that one trace appears under Traces. Ask before you remove Sentry alert rules or CI release steps.

Next steps

Next.js SDK

Replace the Sentry SDK in a Next.js app.

OpenTelemetry

Send errors, traces, and logs from any language.

Set up notifications

Send new issues and incidents to your team.

Read AI triage

Learn what the triage summary tells you.

Get started

Guides

Migrate from Sentry

What moves and what does not

Run both tools side by side

Point the Sentry SDK at Squasher

Issues and logs of the same error

Replace the Sentry SDK

Connect a Sentry organization

Check parity

Turn off Sentry

Give the migration to a coding agent

Next steps

Next.js SDK

OpenTelemetry

Set up notifications

Read AI triage

​What moves and what does not

​Run both tools side by side

​Point the Sentry SDK at Squasher

​Issues and logs of the same error

​Replace the Sentry SDK

​Connect a Sentry organization

​Check parity

​Turn off Sentry

​Give the migration to a coding agent

​Next steps

Next.js SDK

OpenTelemetry

Set up notifications

Read AI triage

What moves and what does not

Run both tools side by side

Point the Sentry SDK at Squasher

Issues and logs of the same error

Replace the Sentry SDK

Connect a Sentry organization

Check parity

Turn off Sentry

Give the migration to a coding agent

Next steps