Fix Guides

Salesforce Froze Our Integration User: Recovery and Prevention Runbook

A frozen Salesforce integration user is a user that Salesforce has locked out, and had its sessions ended, after detecting what it considers suspicious API data access; every sync, job and connected app that runs as that user stops until an admin unfreezes it by hand.

The Symptom: Why Did Salesforce Freeze My Integration User?

Integrations that ran fine yesterday fail on login or authentication, data stops arriving in downstream systems, and the user shows as frozen in Setup. Admins receive an email from Salesforce about it. The exact error text your tools show depends on the tool, so check its own log for the authentication failure and match the time to the freeze.

Salesforce does not disclose what triggers a freeze. Public reporting says enforcement began in sandboxes and production at around 21 October 2026 under Salesforce Help article 005461718, “Core Security Requirement Updates”. The date has moved in secondary sources and we could not load the live article to confirm it, so read the article itself for current dates and wording.

This is different from a daily request cap: if everything fails and recovers on a clock, see Salesforce API limits.

Likely Causes

Because the criteria are not published, these are the conditions that make an integration user look risky, not confirmed triggers.

An integration user with far more access than its job needs — A user on a System Administrator profile, or with View All Data, can read every object. If a connected app or script misbehaves, the amount of data it can reach is exactly what looks suspicious.
A sudden change in volume or pattern — A full re-sync, a backfill, a new report export or a loop that re-queries the same records changes what API traffic from that user looks like.
Credentials or tokens used from somewhere new — A token copied into a new tool, a new IP range, or a credential shared between several systems makes one user look like several.
One user shared across several integrations — When one login backs many tools, you cannot tell which one caused the freeze, and one freeze stops all of them.
No owner and no monitoring — Nobody reads the admin notification email, or it goes to a person who left, so the freeze is discovered when a downstream team notices missing data.

How to Recover — Step by Step

1. Confirm the user is frozen. In Setup, open Users, find the integration user and check whether Unfreeze is offered instead of Freeze. You can also query UserLogin for IsFrozen = true to see every frozen user at once.
2. Read the admin email and the login history. Find the notification Salesforce sent to admins. Then open the user's login history and your connected app usage to see which app, IP and time window preceded the freeze.
3. Work out which integration caused it. If the user backs one integration, start there. If it backs several, pause them and bring them back one at a time. Check for recent changes: a new job, a larger batch, a changed schedule.
4. Reduce what the user can reach. Before unfreezing, move the user to a profile with API Enabled and nothing more, and grant a permission set with only the objects and fields the integration uses. Remove View All Data and Modify All Data unless the integration truly needs them.
5. Unfreeze the user. In Setup, Users, open the user and click Unfreeze. Do this after the previous step, not before.
6. Restore the integration. Re-authenticate or refresh the connection, restart any job that stopped after failures, then check that records written during the outage are not missing. Compare record counts or last-modified timestamps on both sides.
7. Check the whole org for the same exposure. Repeat the permission and ownership review for every other integration user, not only the one that was frozen.

How to Find Which Integration Users Are at Risk

List every user an integration runs as. Include API only users, users who authorized a connected app, and middleware or ETL service accounts.
Check what each can reach. Profile, permission sets, View All Data and Modify All Data, and which objects and fields the integration actually uses.
Check what is attached to each. Connected apps, OAuth tokens and the login history, to spot shared logins and unknown IP addresses.
Find users with no owner. Any integration user nobody can name an owner for is the one you will not notice when it freezes.

Prevention

Least privilege. Give each integration user API Enabled and a permission set scoped to the objects and fields it needs. For connected apps, grant only the OAuth scopes the integration uses.
One integration, one user. A freeze then stops one integration, and you know which.
Restrict where it can log in from. Use IP ranges or login IP restrictions where the integration has a fixed address.
Send the admin notification to a monitored address, not to one person, and write down who owns each integration user.
Watch for it. Alert on UserLogin IsFrozen and on authentication failures in each integration, and review API usage before and after any large job.
Throttle bulk jobs and run backfills deliberately instead of letting them surprise the org.

Find Every Integration User That Could Get Frozen Next

One frozen user is usually a sign that others share the same exposure. Run Wrenk's free health check to find every integration user in your org that could get frozen next.

Related Guides

Salesforce API limits — find which integration is using your API calls.
OAuth token refresh — when a connection needs to re-authenticate.
Continuous health checks — review integration access across the org.

FAQ

Why did Salesforce freeze my integration user?

Salesforce says it can end sessions and freeze users when it sees suspicious API data access. It does not publish the criteria it uses, so the exact trigger for your user is not knowable from the outside. Unusual volume, a new source IP, or a connected app reading far more data than usual are reasonable things to check, but they are hypotheses, not documented rules.

How do I unfreeze a Salesforce integration user?

An admin opens Setup, Users, Users, selects the frozen user and clicks Unfreeze. Admins are emailed when a user is frozen, and the unfreeze is manual. Before you do it, find out why it happened, or the user may be frozen again.

Does unfreezing fix the integration straight away?

It lets the user log in again, but a session that Salesforce ended stays ended. Integrations that cache an access token may need to re-authenticate or refresh, and any that stopped after repeated failures may need a restart.

How do I find integration users that could be frozen next?

List every user that integrations run as (API-only users, users behind connected apps, middleware service accounts), then check each one's permissions, the connected apps and tokens attached to it, and its recent login history. Any with broad data access and no owner or monitoring are the ones at risk.

When did this start?

Public reporting says Salesforce began enforcing this in sandboxes and production around 21 October 2026, citing Salesforce Help article 005461718, 'Core Security Requirement Updates'. That date has moved in secondary sources, and we could not load the live article to confirm it, so treat the date as approximate and check the article for current dates.

Wrenk finds and fixes this across your whole org as part of managed Salesforce operations.

Find Every Integration User at Risk — Free Health Check See Pricing