We have had a lot of issues with spam, whether its cryptovirus emails getting through, or good emails getting improperly blocked. Because of this, we need good message tracing (to find the emails), which we do not feel we have with exchange online. We would like to make the following suggestions:

1. Need to be able to trace further than 7 days back without a 4 hour wait per trace. Our previous message tracing system could go back the entire year nearly instantly, but we need at least 30 days without the 4 hour wait per trace. This was pitched as something that could replace our previous tracing system, and this does not do it.


2. UTC time is confusing in message tracing. It uses “AM and PM.” AM/PM in UTC time is ambigious since UTC time it used all over the world. So either use UTC, or fix the timezone to use our timezone automatically. It looks like very little effort was put into implementing message tracing if this made it past QA.


3. We would like more options in message tracing, especially completely basic fields like searching by subject. I would like to know how this was rolled out as an enterprise product without that a "subject" search field when tracing emails. Part of the reason it is so slow is probably because you have very few search fields to refine.


4. Overall speed improvements, even tracing back less than 7 days or less is slow (sometimes "spins" for several minutes", but more than 7 days is 4 hours+). We basically received a poorly formatted spreadsheet after 4 hours.

A premier case was opened for all of this, the features simply do not exist and its not a configuration issue. They did mention compliance center can help search for emails quicker, but does not trace.

Currently, MFA for Azure AD / O365 is useless regarding protection of mailboxes in Exchange Online, as EWS and MAPI clients can still connect to mailboxes using Basic Authentication, even with Conditional Access rules in place to require MFA, and there's no way of denying this server-side on EXO.

The newly-released Client Access Rules feature promises this functionality in its documentation (see https://technet.microsoft.com/library/mt842508.aspx and https://technet.microsoft.com/en-us/library/dn913650(v=exchg.160).aspx), but unfortunately the functionality is crippled. You can only make rules in the following combinations (info from EXO Engineering team):

OutlookWebApp: BasicAuthentication, AdfsAuthentication
ExchangeAdminCenter: BasicAuthentication, AdfsAuthentication
RemotePowerShell: BasicAuthentication, NonBasicAuthentication
ExchangeActiveSync: BasicAuthentication, OAuthAuthentication, CertificateBasedAuthentication
IMAP4/POP3/OfflineAddressBook/PowerShellWebServices/ExchangeWebServices/REST: none

Note there is no mention of OutlookAnywhere (MAPI) in the above list, but from my own testing, it's impossible to create rules around that protocol. EWS is explicitly mentioned as not possible to restrict auth type for.

IMAP and POP are also bypasses for MFA, but it's easier to turn those off, but many things are dependent on EWS (including Outlook mobile apps), and Outlook (desktop) is of course dependent on MAPI (OutlookAnywhere). It would be useful to control REST too, as I understand some mobile apps will use that.

Yes, you can use CAR's to allow only certain IP's and users to connect with certain protocols, insist everyone VPN's in before connecting, and try to force compliance Outlook client-side, using GPO's to turn on ADAL and then lock down the registry, but this doesn't stop a phisher bypassing MFA for one of your users from whom they've phished username and password, and getting full access to their mailbox.

This is a pretty big security hole for us, and I'm surprised there isn't more noise about it across the internet. I just get blank or stock responses from support and escalation engineers on this, so I'm hoping this post helps.

If you've created a retention label in the Security & Compliance Center and have checked the 'Use label to classify content as a "Record"' checkbox, I would like the ability to delete the label under certain circumstances. If I've never used it, it's not published in any policy, I should be able to delete it. I've set up several "test" labels with this checkbox checked and there is no way (either thru the UI or thru PowerShell) to delete the label definition. Example: if you create a retention label and select the 'record' checkbox, save it and then immediately try to delete it, you will NOT be able to.