OBLYTECH Logo
OBLYTECH Logo
ServiceNow Upgrade Readiness: Guides and Services
Home / Solutions, Guides and Services / Now Platform and Development: Guides and Services / ServiceNow Upgrade Readiness: Guides and Services

ServiceNow Upgrade Readiness: Guides and Services

Upgrade planning is the work of deciding how often a ServiceNow instance moves to a new release, judging what each release will change in your customizations, and setting the go or no-go call before the test cycle starts. This page helps a platform owner set the upgrade schedule, plan release testing and run the readiness checks before each upgrade window.

What an Upgrade Changes on the Now Platform

Each ServiceNow release can change platform behaviour that your customizations rely on, so the upgrade is where those changes meet your code. Upgrades run through Upgrade Center. Its Upgrade Preview tool shows the changes that would affect your current configuration, and the Upgrade History module records, for each upgrade, the changes it applied and the changes it skipped. Staying on a supported release is part of the same decision, so check the support status of your release before you set the schedule.

This page answers questions like: what will break when we upgrade? What new features should we adopt? How do we minimize downtime during upgrades? What is the upgrade schedule that fits our business? How do we manage upgrades across multiple instances?

How Upgrade Planning Relates to Change Management

Upgrades are major changes that require careful planning and testing. Many teams also record the upgrade as a formal change, so that approval is on file before the window opens. Upgrade Center and the Upgrade Preview tool assess impact and readiness, and that assessment is what the change record should rest on.

How to plan the upgrade schedule and release testing

Assess your current upgrade readiness. How frequently do you upgrade (quarterly, semi-annually, annually)? How much downtime do your upgrades require? How many customizations do you have that could break during an upgrade? Do you have a test environment where you can validate upgrades before production? These questions inform your upgrade strategy.

Establish an upgrade rhythm. Some teams upgrade every other release to stay current while reducing upgrade frequency. Plan upgrades in advance: identify the target release, reserve a test environment, schedule a maintenance window, communicate to stakeholders. Use the Upgrade Center and upgrade tools to assess impact, test, and execute the upgrade.

Invest in automated testing. Manual testing of each customization after an upgrade is time-consuming and error-prone. Automated tests (Automated Test Framework) replace manual checks with repeatable ones, so each window spends less time on repetitive testing.

Typical Starting Problems

Organizations delay upgrades to avoid disruption, falling behind on releases and support. They struggle to assess which customizations will break during an upgrade. Testing is manual and slow, extending upgrade windows and increasing risk. Upgrade Center and the Upgrade Preview tool show which changes to review before a window, so the team starts from a list rather than a guess.

What Oblytech Provides

Oblytech helps establish an upgrade cadence and strategy, assess impact of upgrades on your customizations, plan and execute upgrade testing, manage upgrade change processes, and provide managed support during and after upgrades to ensure smooth operation of new releases and quick problem resolution.

Communication around upgrades is critical. Notify stakeholders well in advance of the planned upgrade. Explain what’s changing, how it might affect their work, and what they should expect during the upgrade window. After the upgrade, provide release notes and guidance on new features. Some organizations hold post-upgrade training sessions to highlight new capabilities and best practices.

Planning for recovery is important. If a critical issue is discovered after an upgrade, agree the recovery path in advance, including who calls it and how long it would take. Test the recovery path in advance so you know how long it takes and what needs to happen. Recovery is a last resort, and it has to be planned if something goes seriously wrong.

The Automated Test Framework (ATF) runs the test suites you build, so the same journeys can be run again after each upgrade. Invest in automation to reduce manual testing effort and accelerate upgrade cycles.

Plan upgrades during maintenance windows to minimize impact on users.

Choosing an upgrade rhythm before choosing a date

The first decision is how often the instance will move, because that choice sets the size of every later test cycle. A date on the calendar comes second. Each option below trades the same things differently: how much change each window must absorb, how many customization reviews the team runs in a year, and how far the instance can drift from the vendor’s current baseline before the drift itself becomes the problem.

Rhythm What each window looks like Typical risk to watch
Adopt most releases as they arrive Small change sets, frequent but short test cycles A permanent regression cycle that leaves little time to fix custom code properly
Adopt every other release Larger change sets, one fuller test cycle per window Customization drift between windows that makes each test longer than the last
Adopt only when a deadline forces it One large jump with many changes at once Working outside vendor support while planning, and a schedule set by someone else

Many teams settle on a rhythm they can sustain with the people they have, not the one that looks best on paper. Whatever is chosen, write it down with the reason, so that a future decision to skip a release is a deliberate change rather than a slow drift.

Making the readiness decision in the order it has to be made

  1. Name one person who decides whether the window goes ahead. Shared ownership of a go or no-go call tends to produce a decision nobody can explain later.
  2. Confirm that the target release supports the plugins, connected applications and integration endpoints the instance depends on. Ask the owner of each one, in writing.
  3. List every customization by its business owner, its technical owner and how often it runs. A custom business rule that fires on every incident update needs a different test than a report used once a quarter.
  4. Write the go and no-go criteria before the test cycle starts. Criteria written after the results are known tend to describe what happened rather than what was required.
  5. Book a non-production refresh that is recent enough to reflect real data and configuration, and confirm that the refresh completes within the window the team plans to use.
  6. Decide the rollback trigger in advance. Name the condition, the person who can call it and the point after which rollback is no longer a realistic option.

Signs that the instance is not ready to be upgraded

  • Custom code that nobody has run since the previous release, or whose author is no longer available to explain it.
  • Integrations whose owners on both sides cannot be named, or whose vendors have not confirmed support for the target release.
  • No restore test in the last few upgrade cycles, so the team cannot say how long a rollback would take.
  • High-priority defects in the customization backlog that are being carried into the window as known issues without a workaround.
  • Business events scheduled for the same days as the window, such as month-end close or a large campaign, with no agreed plan for them.

Who does what across an upgrade program

Role Owns Hands off to
Platform owner The rhythm, the go or no-go decision and the rollback trigger Release lead, after the criteria are signed
Release lead The test plan, the schedule and the cutover runbook Support lead, before the window opens
Development lead Custom code results and the fixes they require Release lead, with a list of accepted risks
Integration owners Endpoint checks on each side of every integration Release lead, with evidence attached to the test plan
Support lead User contact, the first-day watch list and escalation Platform owner, at the post-window review

When deferring an upgrade is the right call

Deferral is sometimes correct. A vendor is still reworking a connector your processes depend on, a defect in the target release touches the core request flow, or a business freeze overlaps the window. In each case, record an end date and name the person who reports against it. An open-ended deferral is not a decision, and it usually ends with an upgrade forced by a support deadline and planned by people who have no time to plan it.

For practical guidance on structuring the wider work, the implementation roadmap for upgrades sets out the phases, and the measures that show whether an upgrade program is on course describe what to track between windows. Integration owners will want the integration patterns guide before they confirm endpoint support, and anyone diagnosing a failed window can start with the troubleshooting guide for upgrades.

Frequently asked questions about upgrade planning

Does skipping a release make the next upgrade harder?

It can. Each skipped release adds changes that the next test cycle must cover, and the deltas in release notes stack up. The cost is not linear, because a large jump also brings more deprecated behavior into the same window. If a team skips, it should plan a larger test cycle, not the same one.

Does a small instance still need a full test cycle?

The scale of the cycle can shrink, but the coverage should not. A small instance with a few integrations and little custom code can use a shorter script, provided it still exercises every integration, every role-based approval and the reporting that leadership reads.

Who should be told about a deferral?

Everyone who has a date tied to the upgrade. That includes the business owners of the customizations waiting for fixes, the integration owners who have booked vendor time and the support leads whose rota would otherwise be planned around a window that has moved.

Can feature build work run alongside an upgrade program?

It can, but new build work should not add customizations into the window that is being tested. Freeze the set of changes that will be tested, and queue new build work behind the window so the test results describe a stable target.

The guidance on upgrade planning and operational continuity covers the business side of the timing call.

Upgrade readiness tools, tables and default checks to confirm

Before a ServiceNow upgrade, the records that matter most are the customised platform files recorded in sys_update_xml (some application changes are not recorded there), the results of your Automated Test Framework suites, and the Upgrade History left by earlier upgrades. A readiness review that reads all three has covered the main sources of surprise.

The names below are common out-of-box values. Check them on your own instance and release, because module locations and labels move between releases.

Source Common name What to read before the window
Customer updates sys_update_xml Which platform records your team has changed
System update sets System update sets (feature) Work in progress that could collide with the upgrade
Test suite results Test suite results record Whether the regression suite passed on its last run
Individual tests sys_atf_test Which journeys are actually covered
Upgrade History module Upgrade Center, Upgrade History module What earlier upgrades changed or skipped

Start with the first row, because every customised platform record is a place where the vendor change and your change may meet. Sort the customer updates by the record they modify, group them by owner, and ask each owner whether the change is still needed. Changes that nobody can justify are candidates for removal before the window, which can shorten the later test cycles. The test criteria that follow belong in the testing and acceptance checklist.

Example (hypothetical): a platform owner lists the customer updates and finds a set of changes to a single catalog flow made by three people over several releases. Nobody is sure which still applies. The owners agree to retire two, keep one, and write a test for it in the test framework. The upgrade window then carries one customisation to verify rather than three, and the decision is recorded beside the go or no-go criteria. Teams doing this work as a service can see the scope on platform performance and upgrade readiness.

Where can I see what an earlier upgrade changed or skipped?

Use the upgrade history records and the review of skipped changes in your release’s upgrade tooling (module names vary by release). They list platform files the upgrade did not apply, mostly because a customised version existed. Review those lists first, since each skipped file may hide a fix or feature you have not received.

Which records count as a customisation?

Any record in the customer update table that changes a platform-owned file counts. Entirely new records you created are your own application content, but they can still break when a platform API they call changes, so they belong in the test plan too.

What should an upgrade test suite cover?

Cover the journeys the business cannot go without: submitting a catalog request, opening and resolving an incident, and approving a change. Build them as test suites so the same checks run before and after the window and the results can be compared directly.

More guides on this topic

Judging an upgrade window against the criteria written before testing

An upgrade window is finished when the team can show, criterion by criterion, what held, what broke and which risks were accepted. Book the review date before cutover, so it happens while the people who ran the window still remember what they saw.

Start from the go and no-go criteria, not from the defect list. Each criterion needs one of three answers: met with evidence attached, met with a named risk accepted by a named owner, or not met. A defect that is not linked to a criterion tends to be closed by whoever is busiest, so tie every open item to the requirement it affects before anyone calls it minor.

Watch the first few working days after cutover separately. Support tickets that arrive in that period often trace back to a custom record that passed every test, because the script covered the main journey and not the branch that one team uses at month end. The support lead should keep a short list of those branches, check each one on the first day and again at the end of the first week.

  • Compare the update sets still in progress and the customer update list against the versions taken before the window, and record which changes were overwritten or skipped.
  • Log every rollback trigger that was considered, whether it fired, and who made the call.

Keep the review short and specific. It should produce three things: a decision on whether the release stays, a list of fixes with owners and dates, and an update to the test suite so the next window catches what this one missed. Where a fix was a workaround, write it into the runbook so the next person does not have to rediscover it.

A review that ends in blame rather than changes to the suite will repeat the same problems in the next window. Judge the process first and the people second. If a team wants outside help running this stage, the release governance support service is related reading on release governance.

More ServiceNow guides on this topic

Frequently asked questions

How often should we upgrade ServiceNow?

ServiceNow releases quarterly. Many organizations upgrade every other release (semi-annually) to balance staying current with avoiding upgrade overhead. Plan your schedule based on your needs.

What if an upgrade breaks a critical customization?

Test thoroughly in a sandbox before upgrading production. If a customization breaks, Oblytech helps identify the issue and provide a fix. Have a rollback plan if a critical issue is discovered.

How long does a typical ServiceNow upgrade take?

ServiceNow-managed upgrades involve downtime. Plan a maintenance window with the platform owner, sized to your customization complexity and validation requirements.

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