OBLYTECH Logo
OBLYTECH Logo
Troubleshooting ServiceNow License Optimization: A Practical Guide
Home / Solutions, Guides and Services / IT Asset Management: Guides and Services / License Optimization: Guides and Services / Troubleshooting ServiceNow License Optimization: A Practical Guide

Troubleshooting ServiceNow License Optimization: A Practical Guide

This guide helps a software asset manager trace a bad license figure back to the stage that caused it, whether that is discovery, license data, compliance rules or reporting. It sets out the common complaints in the order to check them, and says when to stop troubleshooting and rebuild a stage instead.

Symptom 1: Software Not Being Discovered (False Negatives)

If software is installed on computers but SAM does not detect it, diagnosis starts with discovery. Is Discovery reaching the computers in question? Check that the MID Server covering each one is running and that the scan credentials are valid there.

Have computers reported discovery data recently? If a computer has not checked in for months, its software is not current in SAM. Are computers behind corporate firewalls that block discovery communication? Some environments block the MID Server connection to the instance, so Discovery reports nothing from those hosts.

The MID Server that covers each network carries that communication to and from the instance, so check the firewall rules on that route. Test with a manual installation: install known software and confirm SAM detects it. If manual installation is detected but production installations are not, the issue might be in how the software is deployed (some deployment methods might hide the software from discovery).

Symptom 2: License Count Not Matching Actual Installations

If SAM shows you have more or fewer licenses than you actually do, the issue is in license data. Are all licenses imported? Check that license acquisition data is complete and accurate. Are legacy licenses tracked? If you acquired a company, are their old licenses in the system?

Are trial licenses tracked separately from production licenses? The gap between what SAM says and reality often indicates incomplete license data. Import all license data, including old licenses and trials.

Symptom 3: Compliance Reports Showing High Non-Compliance

If compliance suddenly shows many violations where previously it was clean, investigate what changed. Did new software get deployed without licenses? Did discovery improve and find previously unknown installations? Did licenses expire and no one noticed? Check whether discovery began reporting software it had missed before.

This can show historical non-compliance that had existed earlier but was invisible. Investigate each major non-compliance and determine if it is a real violation or a detection artifact.

Symptom 4: Cost Data Being Wrong

If cost reports do not match your invoices from vendors, cost data might be wrong. Verify that license prices are entered correctly in SAM. Some vendors give different prices for different quantities or terms; confirm that SAM is using the right price.

Check if discounts are reflected. Some organizations get volume discounts that should be factored into cost calculations. If cost data is wrong, any financial analysis built on it is unreliable.

Symptom 5: Overlapping or Conflicting Licenses

If a computer has multiple licenses for the same software, this is waste. Where to look: run a report of computers with duplicate installations. Determine if both licenses are necessary or if one can be removed. This often happens when merging licenses from different vendors or different business units. Consolidate licenses to eliminate waste.

Symptom 6: False Negatives in Compliance Reporting

If SAM reports that all software is compliant, but you know there are unlicensed installations, the issue is false negatives: SAM is not detecting non-compliance. Where to look: is discovery running on all computers or only a subset? Computers not reporting discovery data are not in SAM, so their non-compliant software is not detected.

Are compliance rules configured correctly? If rules are checking the wrong field or using wrong criteria, they will not flag non-compliance. Test compliance rules against known non-compliant installations to confirm they work.

Symptom 7: Slow SAM Performance and Scalability Issues

If SAM is slow to run discovery, slow to generate reports, or slow to respond to queries, performance is the issue. Where to look: is the SAM infrastructure over-utilized? Check CPU, memory, and disk space on the MID Server host that runs discovery.

Is discovery placing too much load on the client computers it scans? Discovery scans can consume resources, slowing down productivity. Is network bandwidth saturated? Large data transfers from thousands of computers can saturate networks. Solutions include adding server resources, tuning discovery parameters to reduce load, or running discovery during off-hours.

Preventing Common SAM Implementation Mistakes

Prevention is better than troubleshooting. Do not deploy SAM without planning for discovery: scope the effort, estimate timelines, allocate resources. Do not assume license data is complete before importing into SAM: audit license data, remove duplicates, reconcile versions. Do not expect SAM to solve license compliance issues alone: SAM finds problems, but solving them requires process changes (approvals before deployment) and enforcement (removing non-compliant software).

Treat SAM as one tool in a comprehensive license optimization program, not as a silver bullet.

Establishing Effective Support and Escalation

When SAM problems occur, who do you contact? Define escalation paths: frontline support can handle simple questions, but who handles performance issues or discovery failures? Ensure that escalation paths reach people with expertise in SAM, in your infrastructure, and in your organization’s software licensing policies. Create a runbook documenting common SAM problems and troubleshooting steps. A runbook gives support staff a first set of checks to run before they escalate.

Following the data from the device to the report

Many license optimization problems look like reporting problems, and many of them originate further upstream. A compliance figure that looks wrong has passed through at least four stages: a device reports what it has installed, that installation is matched to a software product, the product is matched to an entitlement and the report applies a rule to the result.

A fix applied at the report stage will hide the fault without removing it. Written for the administrator tracing a bad figure back to its origin, this part gives a method that works whichever stage raised the complaint.

A diagnostic table for the most common complaints

When a complaint arrives, classify it by what the person actually observed, then check the stages in the order shown. The second and third columns are the checks; the last column is when to stop and escalate instead of continuing.

What was observed First check Second check Escalate when
A product is missing from the estate Whether the device has reported within the normal cycle Whether the product name is recognized by the matching rules A whole network segment is absent from discovery
Installed count is higher than purchased Whether one physical installation has two records Whether a virtual machine clone was discovered as a separate host Duplicates recur after each discovery cycle
Cost figure differs from the invoice Whether the price on the record matches the order form Whether the price tier matches the quantity band in the contract The vendor’s invoice itself is inconsistent with the contract
Renewal shows a sudden jump Whether a retirement or a reduction was entered with the wrong date Whether a bundle was reallocated to this product The vendor changed its metric without notice
Report is slow or times out Whether the report runs against all history rather than a bounded period Whether scheduled jobs overlap with discovery Performance returns after each schedule change

A method that avoids chasing the symptom

  1. Write down exactly what the person saw, including the report, the date and the filter applied.
  2. Reproduce it on the same report, not on a fresh export that may behave differently.
  3. Pick the record that best represents the problem and follow it upstream through each stage in the table.
  4. Note the first stage where the record disagrees with what you know to be true. That stage is the cause to fix.
  5. Correct the source record or the matching rule, not the report output.
  6. Re-run the original report and the test cases that touch the same stage, so that the fix does not move the fault elsewhere.

Errors that repeat after they appear fixed

Three patterns commonly cause repeat tickets. The first is a manual correction on a record that the next discovery run overwrites, so the error returns each cycle. The fix is to correct the upstream source or a matching rule, and to log the override so it can be reviewed.

The second is a rule that matches a product name too broadly, catching an unrelated product with a similar label. The third is a scheduled import that runs after a manual change and quietly reverts it. Check the job order before blaming the data.

The discovery planning post covers how to keep scope tight enough that these collisions are less frequent.

Credentials and coverage gaps that look like license faults

Discovery that cannot log in to a device will often record less than the device holds, and the license report will then look as if the product is under-deployed. Before concluding that a product is unused, confirm that the credentials used for discovery were valid on the devices in question and that the scan reached them.

Credential governance, including who may change those credentials and how rotation is handled, is covered in the discovery credential governance post. A coverage gap of this kind should be recorded as a known limitation on the report, so that readers do not treat an absence as evidence of non-use.

When to stop troubleshooting and rebuild a stage

Some faults are not worth diagnosing record by record. If the matching rules have accumulated years of exceptions, if the product catalog contains many near-duplicate names, or if the entitlement records were loaded from three different spreadsheets with no common key, repeated fixes will cost more than a controlled rebuild of that stage.

Rebuild in the order the data flows: catalog first, then matching, then entitlements, then reporting. Rebuilding the report alone will not help, because the report is only as correct as the stages beneath it.

Frequently asked questions about troubleshooting

How do we know whether a fault is in the data or in the report?

Take one record and read it at each stage. If the record is correct in the source but wrong in the report, the fault is in the report logic. If the source is already wrong, the fault is upstream.

Should we keep a log of every correction?

Yes, and the log should record the record, the stage, the previous value, the reason and the person who made the change. A log of corrections is often the quickest way to find a rule that is systematically wrong.

Who should be called in first when the fault is unclear?

The person who maintains the matching rules, because unclear faults often turn out to be a rule applying to a case its author did not anticipate. Involve the discovery administrator when coverage is in doubt.

Is it safe to correct compliance figures before an audit?

Only if each correction is traceable to a source document and the original figure is preserved. A correction that cannot be explained is likely to look worse to an auditor than the original fault.

What should be done when the vendor’s own data disagrees with ours?

Record the disagreement with both figures and their sources, then raise it through the vendor manager. Do not silently adjust the internal record to match the vendor’s figure until the difference is understood.

Next step

A recurring fault is cheaper to fix at the stage where it starts than at the report that displays it.

Software license optimization in ServiceNow: table names and values to check first

Software license optimization compares what is installed with what the organisation has bought. The deciding fact is that the two sides sit in different records: installations in cmdb_sam_sw_install and entitlements on the license record, so a wrong figure usually traces to one of those two sources or to the product model that joins them.

These are common out-of-box names, so check them on your own instance and release, because labels, plugins and defaults change between releases.

Name What it holds First check
cmdb_sam_sw_install Typically one row for each software installation found on a computer Count rows for one product on one device to spot duplicates
Software Model form The record that one or more discovery models are matched to, representing the version or configuration you buy or hand to users Confirm the installation resolved to a software model through its discovery model
Software license record The software license (entitlement) record, including purchased rights Compare it with the purchase order
MID Server The Java application, running as a Windows service or UNIX daemon on a server in your local network, that covers the device Confirm the server covering the device is running

Example (hypothetical): a software asset manager sees a design application reported with more installations than entitlements. Following the records upstream, the manager finds that several installations belong to cloned virtual machines that discovery recorded as separate hosts. The gap is a counting artefact in the installation table rather than a purchase shortfall, so the correction belongs on the host records and the report stays untouched.

For the wider picture, the software license optimization guides cover that topic in more depth, and the software asset management professional section covers its neighbouring product. Data feeds into the module are covered under license data integration patterns.

Which table holds the number of licenses an organisation owns?

Entitlements sit on the software license record, which typically carries the purchased rights and the metric used to count them. Installations are held in a different table, so the two do not share a row. Compare them through the product model instead of searching for one table that holds both.

Why does one product appear under two names in the totals?

Normalisation standardises the publisher and product names on each discovery model, and matching then links that model to a software model. When a discovered name does not match any model, the installation typically stays unmatched, so it may be missing from the count for the product you expect, and the same product can look split across entries. Correct the model or the matching, rerun the process, and only then compare counts.

Can a MID Server fault change a compliance figure?

Yes, and it can go unnoticed. If the MID Server covering a network cannot reach its devices, their installation data stops refreshing, so the compliance figure can age without an obvious sign. Check the MID Server that covers the device and the last discovery date on the device before suspecting the compliance rules.

How to confirm a software license fix by comparing the report before and after the change

A fix to a software license figure is only confirmed when the same report, run with the same filter, moves where you expected and nowhere else. Before you change a discovery setting, a product match or a compliance rule, export the report and keep the file with its date and filter. That export is your baseline. Without it, you cannot tell whether the figure moved because of your change or because a discovery cycle ran in the meantime.

Change one stage at a time. If you alter a product match and a compliance rule in the same afternoon, a shift in the total will not tell you which change caused it. Apply the change, wait for the next full discovery cycle on the affected devices, then run the report again with the same filter.

Read the difference by product, not by total. A correct fix moves the products it targets and leaves the rest of the list alone. Check the result against these points:

  • Did the targeted product change by the amount you expected, and no more?
  • Did any unrelated product move? If so, the match or rule is too broad.
  • Did the change survive the next cycle? A figure that returns after discovery can mean the correction sits in a record that discovery overwrites; related background is in duplicate control for CMDB records.
  • Did the report total change while no single product did? Look for a grouping or rounding rule applied at the report stage.

Keep a short change record next to the baseline export. Note the stage you changed, the reason, the date and who approved it. When an audit asks why a figure moved, the record answers it without a search through old tickets. Rollback is also simpler: if the comparison shows the wrong products moved, you reverse the single change and return to the baseline instead of guessing which of several edits caused the damage.

Stop comparing and rebuild the stage when every test shows movement you cannot explain. That usually means the stage holds conflicting definitions, and more small edits will only add to the confusion. For a structured review of the stages before you decide, the software asset compliance workflows page sets out how each stage should hand over to the next.

Frequently asked questions

How often should we run license compliance audits?

At least quarterly. Quarterly audits balance cost with the pace of software changes. Very rapidly changing environments might audit monthly. Make sure discovery is running frequently enough to feed audits with current data.

What should we do if a vendor audit finds unlicensed software?

Immediately license the software or remove it. If the violation was accidental (software installed without approval), license it and investigate how it got installed. If it was deliberate (budget constraints), prioritize licensing it in the next cycle. Document the violation and your remediation step. This shows good faith to the vendor.

How do we prevent license non-compliance from happening again?

Establish controls. Require approval before software is deployed. Periodically reconcile installed software against approved licenses. Educate users about license compliance. Use SAM to monitor and alert on new non-compliance so you catch problems quickly rather than waiting for a vendor audit.

Talk to Oblytech

Planning, running or fixing a ServiceNow programme? Tell us where you are and we will suggest a sensible next step.

Contact us