System Advisor

The System Advisor reviews your system's configuration and security posture and reports any issues it finds, along with a plain-language explanation of what each issue means and how to fix it. It replaces the older "System Alerts" list. You'll find it under Admin > System Configuration > System Advisor, and the admin menu shows a badge with the number of open issues that need attention. The page is available to full administrators only — an administrator restricted to particular areas will not see it in the menu.


Health scores

At the top of the page are two dials, one for Security and one for Configuration, each showing a score out of 100 for that area. The score answers a different question from the issue list below it: the list tells you what is wrong, the score tells you how much of what was checked is actually in good shape.


Each score is the share of that area's checks that passed, weighted so that more serious checks count for more than minor ones. Underneath each dial is the plain count it is based on, for example "6 of 8 checks passing". Checks that do not apply to your system are left out of the score entirely rather than counted as passing — a hosted system is never scored on the on-premise email and document storage settings, an on-premise system is never scored on database usage, and a site that is not licensed for AI is never scored on the AI checks.


The color of each dial is a quick verdict. Green means 90 or above, yellow means 70 to 89, and red means below 70. There is one override: if the area has even one open Critical issue the dial is red no matter how high the score is, because a single wide open hole is not a good result with a caveat. A dial only reads 100 when every check in that area has passed outright, so a score can sit at 99 while a single small item remains.


Click a dial to show only that area's issues and hide the other. Click the same dial again, or the "show all" link beneath the dials, to go back to showing everything.


The two sections

Findings are grouped into two stacked sections:

  • Security — issues that affect who can access the system and your data (shown first, because they matter most).
  • Configuration — correctness, hygiene, and performance issues (missing settings, record-type problems, storage cleanup, and so on).


Each section shows a count of its open issues. A green check and "No open … issues" means that section is clean. Each section can also be shown on its own by clicking its dial at the top of the page.

Severity

Every issue carries a severity:

Severity

Meaning

Critical

Should be addressed promptly — affects security, compliance, or core functionality.

Warning

Worth addressing — a gap or risk that isn't immediately breaking anything.

Low

Minor, but still counted in your score — housekeeping or a small gap worth reviewing.


Within each section, issues are ordered Critical, then Warning, then Low.


Working with an issue

Each issue is a compact one-line row. Click it to expand and you'll see:

  • Why it matters — a short explanation of the impact.
  • How to fix — concrete remediation guidance.
  • Go to fix — jumps you straight to the exact admin screen (and, where relevant, the exact settings panel) where the issue is resolved.
  • Learn more — opens the help topic that covers the setting behind the issue, when there is one.


Scan for Changes

The list reflects your system's current state every time you open the page. The Scan for Changes button (top right) runs an authoritative scan: it records what's currently found and automatically marks anything that's no longer a problem as resolved. Use it after you've fixed something to confirm the issue clears and to add it to the Recently fixed history. Opening the page keeps the list up to date but never marks anything as resolved, so after fixing something you will still want to press Scan for Changes to move it into Recently fixed.


Acknowledging issues

Some findings are real but intentional for your environment — a record type that is deliberately public, or an AI setting your organization has thought about and chosen. Rather than have the Advisor flag them forever, you can Acknowledge an issue:

  • Expand the issue and click Acknowledge.
  • Optionally enter a short reason (recommended — it documents why the issue is accepted).
  • The issue moves into a collapsed "Show acknowledged (N)" strip within its section, and the open-issue count (and menu badge) drops.


Acknowledged issues are never hidden completely — they remain visible (collapsed) so they can be revisited, and each shows who acknowledged it and when. To bring one back to the active list, expand it and click Un-acknowledge.


Acknowledgements are tracked per issue, so accepting one item (for example one public record type) does not suppress others of the same kind that appear later.


Seeing what passed

Each section has a collapsed Show passing checks (N) strip, above the acknowledged strip. Expand it to see the checks in that area that are not reporting a problem. This is what makes the score meaningful — without it, a score of 83 is just a number, and there is no way to see what the Advisor actually looked at.


Two kinds of entry appear in this list. A green tick means the check found nothing at all. A grey tick followed by "accepted" means the check did find something, but you have acknowledged every one of its findings — you have seen it and signed off on it. Both count towards the score, but they are shown differently because "nothing is wrong" and "I have accepted this" are not the same statement.


Acknowledging an issue improves the score. Where a check reports several findings at once, such as one per Record Type, acknowledging some of them gives you partial credit rather than nothing, so working through a long list is rewarded as you go. That check only moves into the passing list once every one of its findings has been acknowledged or fixed. This is why fixing one exposed Record Type out of six clears that one row from the list but does not yet turn the check green.


Because acknowledged findings count as passing, the "N of M checks passing" figure under a dial can differ slightly from the number of green ticks in the list. Both are correct — the dial counts a fully acknowledged check as passing, and the list shows you which of those are genuinely clean and which are accepted.


Recently fixed

The Recently fixed panel is an embedded history of issues that have been resolved. Each entry shows the issue and who resolved it and when (the administrator whose scan detected the resolution). It's a lightweight record of what's changed over time — expand the panel to review it.


Note: simply opening the System Advisor page records the issues it currently finds, so an issue you fix later can still be reported as fixed. Marking it fixed still requires a Scan for Changes — opening the page never resolves anything on its own.


Alerts on a Record Type

The Advisor's Record Type checks also run on their own whenever you open a single Record Type for editing under Admin > Record Types. If that Record Type has any findings, an Alerts section appears at the top of the editor with a count, opened for you, and the General Settings section starts collapsed behind it so the alerts are the first thing you see.


Only the checks that concern one Record Type appear there: its filters, the filter and default value of its fields, duplicate filters it shares with another Record Type, its field and category counts, and the Records of that type sitting in the Recycle Bin. System-wide findings such as MFA or the Base Server URL only ever appear on the System Advisor page.


This panel is a read-only heads-up rather than a working list. It shows each finding's severity and headline text, colored the same way as the System Advisor page, and there is nothing to expand, acknowledge, or resolve. The full explanation, the fix guidance, and the Acknowledge button for these same findings are all on the System Advisor page.


Security checks

Check

Severity

What it means / how to fix

'ADMIN' user is still active and should be deleted, renamed, or disabled

Critical

The default ADMIN account is a well-known login and a common attack target. Delete, rename, or disable it once your own administrator accounts are in place.

'Base Server URL' is not secure (http)

Warning

The Base Server URL uses plain http, so logins, cookies, and shared links travel unencrypted. Serve the site over https and update the Base Server URL.

Multi-Factor Authentication is not enabled

Warning

Accounts are protected by a password alone. Enable MFA under the MFA settings on the Main Options screen and choose at least one method.

Password policy is weak or not set

Warning

The password complexity requirement is at its lowest level. Raise the Password Security level (Medium or higher) under Security settings.

Session inactivity timeout is not set

Warning

Unattended sessions can stay open indefinitely. Set the inactivity timeout under Security settings.

Session inactivity timeout is very long

Low

The timeout is set to more than 8 hours, so unattended sessions can stay open for an extended period. Lower it under Security settings.

Base Security gives every user read/write access by default

Warning

The system-wide Base Security level grants all users read/write to records unless a Record Type restricts it. If unintended, lower it and grant access per Record Type or Team.

Record Type is exposed to anonymous users

Low

A Record Type is reachable by anonymous/external users (public access or collaboration). Confirm the exposure is intended and only appropriate data is shared.

Administrator has not logged in for over 6 months

Warning

An unused administrator account widens the attack surface. Confirm it's still needed; disable or remove it if not.

Public E-form pulls data from a Record Type that is not marked public

Warning

Only appears while Secure API Mode is OFF. A public E-form reads records from a Record Type that has no public setting turned on, so outside users can see data from an otherwise private Record Type. Confirm the data is safe to expose, then turn on Secure API Mode and turn on Public Data Pulls for that Record Type. One finding per form and Record Type pair.

Public E-form pulls data from a Record Type that is blocked

Warning

Only appears while Secure API Mode is ON. The pull is being blocked, so the lists, charts, calendars or maps it feeds are empty for public users while still working for logged in users. Turn on Public Data Pulls for that Record Type, or remove the data pull from the form. One finding per form and Record Type pair.

MCP server is enabled

Warning

AI clients can connect to this tenant at /mcp using a DocMgt user's credentials and act with exactly that user's rights. It is off by default, so it was switched on deliberately at some point. Confirm the access is still wanted and that it connects as a dedicated, restricted user rather than an administrator; otherwise turn off Enable MCP Server on the AI Settings screen. Acknowledge it to record that it is intended.

Configuration checks

Check

Severity

What it means / how to fix

End User Agreement not Accepted

Critical

The End User License Agreement hasn't been accepted. Accept it on the Software / License Info screen.

'Base Server URL' is not set

Critical

Generated links (emails, shares, portal) can't be built without it. Set it on the Main Options screen.

'Base Server URL' does not match your current URL

Critical

The configured URL differs from the address you're using; generated links may point to the wrong place. Update it to match.

Database Usage High (hosted only)

Critical

Average or peak database CPU is above 80%. Sustained high usage can slow the whole system — review recent activity and scheduled jobs, and consider scaling up.

Email Settings Not Configured (on-premise)

Warning

SMTP is incomplete, so notifications, shares, and workflow emails can't be sent. Enter the SMTP server, port, and From address.

Document Storage location Not Set (on-premise)

Warning

File content is being written into the database, which bloats it and hurts performance. Set a document storage location.

WorkTrigger has a duplicate GUID

Critical

Two WorkTriggers share the same GUID, which can confuse workflow steps. Re-create or re-import one so each is unique.

Record Type has no filter set

Warning

Without a filter, the Record Type can't be matched automatically when records are routed or classified. Set a filter value on at least one field.

Record Type field filter does not match its default value

Warning

New records may not match the Record Type's filter. Align the field's default value with its filter value.

Record Type has duplicate filters

Warning

Two Record Types share identical filters, so incoming records can't be reliably classified. Make the filters unique (or merge the types).

Record Type has too many defined fields

Warning

A very large field count is hard to maintain and can slow data entry and search. Remove unused fields (aim for 25 or fewer).

Record Type has too many defined categories

Warning

An excessive category count clutters the UI and can degrade performance. Consolidate or remove categories.

More than 5000 Records in Recycle Bin

Low

A large recycle bin consumes storage and slows admin operations. Review and permanently purge old records.

Oldest recyclable Record is more than 12 months old

Low

Records have sat in the Recycle Bin for over a year. Purge what you no longer need.

Public E-form shares its public URL name with another form

Warning

Two public E-forms use the same public URL name, so a link of the form /public/form/<name> resolves to an arbitrary one of them. The wrong form can be served, and browser "Save for Later" drafts collide across the two. Give one of them a unique Public URL name, or clear it so that form is reached only by its numeric ID.

Workflow AI has no monthly token limit

Critical

The Workflow AI Limit is set to unlimited, so every token the AI workflow actions consume is billed with no ceiling. Automation usage scales with how many documents you process rather than with how many people use the site, so a workflow that misfires, or one oversized batch, can run up a very large bill before anyone notices. Set a monthly ceiling under Workflow AI Limit on the AI Settings screen, or acknowledge this if your organization has accepted uncapped Automation billing.

AI chat transcripts are kept forever

Warning

No retention period is set for AI Chat, so every transcript is kept indefinitely. Transcripts hold the questions your users asked and the record data the AI was given to answer them, and they accumulate for as long as people use the feature. Set a retention period for Chat under End User AI Retention on the AI Settings screen, or acknowledge this if keeping the full history is deliberate.


The last three checks — the two above and MCP server is enabled — only run on sites licensed for AI. On any other site they are skipped entirely and play no part in either score, because the settings they look at cannot be reached there. All three are worth a word about acknowledgement: none of them is wrong on its own. Each one is a decision someone should have made on purpose rather than arrived at by leaving a default alone, so acknowledging one with a reason is a perfectly good outcome and counts as a pass.


Tips

  • Open the page periodically — the menu badge tells you at a glance how many open issues there are.
  • Use Acknowledge (with a reason) for accepted exceptions instead of ignoring them, so the list stays meaningful and the decision is documented.
  • After fixing several items, click Scan for Changes to clear them and build up your Recently fixed history.
  • Expand Show passing checks in either section to see what the Advisor examined and found to be in good order — useful when you want to know what a score is based on.