Software maintenance: what a monthly plan should include
Work, priorities, response times and scope to clarify before handing over an application’s ongoing care.

A software maintenance plan should specify which application it covers, what work is included, how requests are handled and who checks the changes. The monthly fee tells you little about how the application will be kept running.
Compare plans by giving providers concrete situations. An error blocks the team's work, a dependency needs an update, an external service changes or someone requests a feature. Putting everything under “support” makes the response and the cost hard to predict.
Identify the application and its components first
The scope should identify environments, components, integrations and responsibilities. A plan covering the frontend may assign the database, infrastructure and third-party services to other owners.
Ask the provider to document the starting conditions too. Include the software version, available access, known issues and the limits of the plan. A stalled or fragile project may need a separate proposal for initial recovery work.
Fixes, updates and new features
Group requests into these areas to compare plans. Providers may use different terms, so compare the work described and the limits in writing.
| Area | Example | Question to ask |
|---|---|---|
| Fixes | An agreed export produces incorrect data | Which bugs are covered and what limits apply? |
| Updates | A library or integration needs a new version | Does the plan cover analysis, testing and release? |
| Routine checks | Review errors, availability or backups | Which checks are done and who receives the results? |
| Enhancements | A new role, report or approval step | Is there an allowance or does this require a separate estimate? |
| Exceptional work | Migration or replacement of a major component | How do we authorize work and costs outside the plan? |
Even a short request can have a wide impact. “Add an approval” can affect permissions, notifications, data and reporting. The plan should explain how the provider assesses the impact before work starts.
Updates also need verification
In a GitHub repository, Dependabot can suggest dependency updates through pull requests. The team still decides which changes to accept and how to verify them in the application. GitHub Docs, Configure Dependabot version updates.
In npm projects, npm audit checks the dependencies in scope for known vulnerabilities. Review alerts and agree which tests accompany updates, including checks for permissions, data and application logic as relevant. npm Docs, npm audit.
Before a release, identify the workflows that the change may affect and the cases that will be tested. Include those checks when describing the work.
Response time and resolution time measure different things
An initial response confirms that the report was received or assessed. Resolution requires reproducing the issue, understanding it, making a change and verifying the result. Missing access, external services or effects on data may determine how long that takes.
An agreement should specify coverage, contact channels, priorities and the conditions for any guaranteed times. Check whether evenings, weekends and holidays are covered. Put the agreed availability in writing.
| Situation | Information needed for assessment |
|---|---|
| Essential workflow blocked with no workaround | Which activity has stopped, who is affected and when it began. |
| Feature degraded, workaround available | What work is still possible and which limits the team faces. |
| Issue limited to one case | How to reproduce it and whether it involves data or integrations. |
| New requirement | Expected outcome, reason for the change and people affected. |
Use these examples to classify requests. Agree on priorities for the way your company works before the first urgent issue arises.
Monitoring and backups: ask what gets checked
An availability check can show that a service responds. To know whether orders still go through, ask which workflow signals are observed, which failures trigger alerts and who is responsible for responding.
A backup plan should say what is backed up, how long it is kept, who is responsible and how to recover it. PostgreSQL documentation, for example, describes the dump and the steps needed to restore it. Ask whether the plan includes a test restore, in which environment and against which criteria. PostgreSQL, SQL Dump.
If another provider manages hosting or the database, agree on how to involve them. Ask who holds the files, who can restore them and which parts of the application they cover.
How to handle requests and reporting
- Record what should happen, the steps that reproduce the issue and its effect on work.
- Have the contact confirm priority and whether the request is covered or needs an estimate.
- Verify the change with the affected cases and release it where agreed.
- Tell the contact the outcome and any limits that remain.
- Use a regular summary to show completed work, open issues and decisions needed.
If the plan reserves hours or capacity, clarify how analysis, meetings, testing and releases are counted, what happens to unused capacity and how additional work is approved. For other billing models, ask what the plan covers and what needs a separate proposal.
What determines the cost of software maintenance
Cost depends on the application, its starting condition, integrations, the importance of its workflows, the coverage requested and the work agreed. Unclear access or manual release steps can create initial work that should be priced separately from recurring service.
Give providers the same inventory and use cases so their proposals are comparable. Ask them to break down the monthly fee, setup, third-party services and exceptional work. A percentage of the original development cost leaves those distinctions out.
A checklist before choosing a plan
- Does the plan identify the application and distinguish third-party services?
- Does it list included tasks, limits and exclusions?
- Does it state coverage and priorities, with separate definitions for response and resolution?
- Will the team know how testing, releases and recovery are handled?
- Who reports, approves and verifies the work?
- Who approves extra work before it starts?
- Will the company retain the access and documents it needs for a future handover?
That last point matters even when the relationship is going well. Whoever looks after the application needs enough information to keep it running. For more on accounts and materials, see the guide to digital product ownership.
Frequently asked questions
Does a monthly maintenance plan cover all new features?
Can I ask for bug fixes alone?
Can someone who did not build the software maintain it?
Who will look after the software your team uses every day?
If you want to clarify how bugs, updates and new requests will be handled, we can assess whether LuCz can support the application with a monthly maintenance plan. Tell us which application you use, which problems recur and who looks after it today.
In an initial call, Matteo reviews the context, fit and possible first assignment. Any initial work and monthly activity are defined for your case, with agreed responsibilities and coverage.
On the booking page, enter your contact details and a short description, then choose a day and time. Any technical assessment, access or subsequent work is agreed separately, with the scope, responsibilities and cost made explicit.
Sources
- GitHub Docs, Configure Dependabot version updates.
- npm Docs, npm audit.
- PostgreSQL 18, SQL Dump.
- LuCz, Stabilization and maintenance for existing software.
- LuCz, Digital product ownership.
Sources were consulted on November 2, 2026. The classifications, examples and checklist are proposed criteria for evaluating an agreement. Check the actual agreement for service coverage.
Written by

Matteo Lucrezio
Startupper | Lead Software Engineer
Book 30 minutes with Matteo
Leave your contact details, then choose a day and time for a free 30-minute initial conversation. Any technical assessment and subsequent work are agreed separately.
No suitable time or prefer to write? Email me at matteo.lucrezio@lucz.dev.