There are two sets
Living Security reaches your provider over two different networks, and they serve different parts of the integration.
Most integrations need only the sync set. A specific group needs both, permanently — see Integrations that need both sets.
Integration sync traffic
These four are the same for every Living Security region. Customers hosted in the EU use the same four as customers hosted in the US.
For providers that accept a comma-separated list:
Platform traffic
Unlike the sync addresses, these are region-specific. Use only the pair for the region that hosts your organization. If you are not sure which region that is, ask your Customer Success Manager before you configure anything — allowlisting the wrong region’s addresses leaves the integration failing in a way that looks like it has nothing to do with the allowlist.
For providers that accept a comma-separated list:
Outbound playbook actions
By default, playbook actions that write back to a connected system route through the same path as your data syncs, so they leave from the sync addresses. Nothing extra is needed. An administrator can switch an individual connection to direct egress, which moves that connection’s actions to the platform addresses for your region. That setting exists for exactly this situation — a provider instance that restricts API access by source IP and needs a small, stable, Living Security-owned set to allowlist. A few providers always use direct egress because they cannot be proxied at all. If you allowlist only the sync addresses and someone later switches a connection to direct egress, that connection’s actions start failing while its syncs keep working.Integrations that need both sets
For most integrations, Living Security’s integration service obtains and renews the access token itself, and everything reaches your provider from the sync addresses alone. For the integrations below, that service cannot renew the token, so the Living Security platform mints it directly — once when you connect, and then continuously for as long as the connection is active. Those token requests leave from the platform addresses. The scheduled data syncs still leave from the sync addresses.- Auth0
- Cofense Triage
- Code42
- Cornerstone
- CrowdStrike Falcon
- Microsoft Entra ID — including Microsoft Graph
- OneLogin
- Sophos Central
- Zscaler Client Connector
Add every address in the sets you need
Traffic can leave from any address within a set, and which one is used for a given request is not fixed. Allowlisting a subset produces intermittent failures that are difficult to diagnose: some requests succeed, others fail, and the pattern looks unrelated to configuration. This applies to the platform pair as much as the sync four. Each region runs two, and requests land on either.Which integrations need this
An IP allowlist is a setting in your provider, not in Living Security. It applies only if your organization has enabled that restriction. If you have not enabled IP restrictions in the provider, no action is needed. These providers commonly require it:
Other providers may offer a similar setting. Check the provider’s page in the Catalog and the provider’s own documentation.
Symptoms of a missing or outdated allowlist
The two sets fail in opposite ways. Match the symptom to the set before changing anything.The connection succeeds, but data never arrives
This is the signature of a missing sync address. Credentials are validated at connection time over a path that is often not subject to the same restriction, so the connection dialog in Living Security reports success. The failure appears later, when a scheduled sync runs and is blocked. Look for:- The connection shows as connected, but streams report no records.
- Sync runs fail with
401 Unauthorizedor403 Forbidden. - An integration that worked for months begins failing with no change on your side.
The connection cannot be created at all
This is the signature of a missing platform address, and it applies only to the integrations listed in Integrations that need both sets. Look for:- Setup fails while obtaining an access token. No connection is created.
- The error reports that a host could not be reached, with no HTTP response behind it.
- The provider’s own sign-in or Conditional Access logs show a blocked request from an address you have not allowlisted.
A connection that cannot be reached at all is not a credential problem. No authentication took place, so rotating the secret or regenerating the client credentials will not help. An unallowlisted source address is the most common cause, though a wrong host name, a wrong port, or a provider-side outage look the same from our side.

