TLDR overview
- Mean time to resolve (MTTR) is an operational metric that shows how long, on average, your organization takes to resolve an issue after detecting it, and whether that speed is improving over time.
- DORA's Articles 11, 12, 13, and 17 ask financial entities to resolve incidents quickly and recover critical functions within the objectives they set themselves, which is where a well-measured MTTR becomes useful supporting evidence.
- Because the average only counts closed items, it can improve while your riskiest issues stay open, so it belongs next to open-item age and overdue critical findings in any report you show an auditor.
- SonarQube tracks MTTR for code findings, which gives you a solid piece of code-remediation evidence, while production-outage recovery and the wider DORA evidence package stay with your incident and resilience processes.
Financial entities operating under the Digital Operational Resilience Act (DORA) need to show that they can respond to, recover from, and learn from ICT-related incidents, and measurement is one of the most practical ways to make that visible. Mean time to resolve (MTTR) is often one of the first metrics teams reach for, because it is familiar and easy to track over time, although it is also easy to misread when the definitions behind it are not clear.
This page explains what MTTR measures, which DORA obligations depend on how quickly you resolve and recover, and how to keep the metric meaningful as evidence. It is written for the security, compliance, and resilience owners who connect regulatory obligations to control evidence, and for the engineering, SRE, and remediation teams who produce that evidence in practice.
What does DORA require for incident response and recovery time?
What MTTR is
Mean time to resolve (MTTR) is the average time between detecting an issue and resolving it, calculated across the items resolved within a defined reporting period. You can apply it to production incidents or to code findings, and each of them deserves its own calculation, since they have different lifecycles and owners. Used consistently, MTTR shows how quickly your organization closes out problems and whether that speed is improving over time.
What DORA requires that depends on time to resolve
DORA (Regulation (EU) 2022/2554) describes its obligations in terms of capabilities and outcomes, and several of them depend directly on how quickly you resolve incidents and recover affected functions:
- Response and resolution (Article 11). Financial entities must maintain an ICT business continuity policy with documented arrangements to respond to and resolve ICT-related incidents quickly and effectively, limit damage, and prioritize the resumption of activities.
- Recovery objectives (Article 12). Entities must maintain backup, restoration, and recovery procedures, and set recovery time and recovery point objectives for each function based on its criticality.
- Learning after incidents (Article 13). After a major incident disrupts core activities, entities must run post-incident reviews to analyze root causes and identify improvements.
- Incident management (Article 17). Entities must run a defined process to detect, record, classify, and follow up on incidents, so that root causes are addressed and recurrence is prevented.
How MTTR and DORA work together
DORA describes the outcomes it expects, from quick incident resolution to recovery within the objectives you have set, and MTTR is one of the ways you can show whether those outcomes are actually happening. Tracked over time, it gives you a trend line for the prompt resolution that Article 11 asks for and a reference point you can hold up against the recovery objectives you defined under Article 12. If the post-incident reviews under Article 13 and the root-cause follow-up under Article 17 are working, you should also see recurring issues gradually drop out of the data.
That is what makes MTTR valuable for compliance teams: a well-measured MTTR, backed by a clear evidence trail, strengthens your case that your response and recovery controls are working as intended.
MTTR, recovery objectives, and reporting deadlines are different things
Several timing concepts sit close to MTTR and are often confused with it, although they measure different things and carry different obligations, so it helps to keep them clearly apart.
Resolution time and service restoration time
Incident resolution time covers the full period until an incident is resolved for good, while service restoration time stops the clock as soon as the disrupted function is running again, even if the underlying cause is still waiting for a permanent fix. The difference matters for DORA, because Article 17 requires root causes to be followed up and addressed, and an incident can be marked as restored while the defect that caused it is still open.
Code remediation time
Code remediation time measures how long a code finding stays open between detection and resolution. It belongs to the software development lifecycle, which makes it a useful input to your secure development evidence, and it should be kept apart from production incidents, since averaging the two into one number produces a figure that is hard to interpret.
Regulatory notification deadlines
DORA also sets incident notification and reporting obligations toward competent authorities, and these run on their own timeline. They govern when you have to inform the authority about a major incident and are tracked separately from how quickly you resolve it internally. The exact reporting windows are defined in the accompanying technical standards and are outside the scope of this page.
How do you calculate MTTR?
You calculate MTTR by adding up the time between the start and stop events for every item resolved in a reporting period and dividing it by the number of those items. The result is only as reliable as the definitions behind it, so agree on these parameters and document them before you report the number:
- Start event: Decide when the clock starts, which is typically detection or the creation of an incident or finding record.
- Stop event: Decide when the clock stops, which should be verified resolution rather than provisional restoration.
- Reporting period: Fix the window you calculate over, for example the last 30 days, and keep it consistent between reports.
- Resolved population: Define which items count, and in particular whether open items are included or excluded.
- Severity and service criticality: Segment by both, so a critical outage and a minor finding never end up blended into the same average.
Here is a hypothetical example with invented numbers. Say that over a 30-day period your team resolved six incidents, which took 2, 3, 4, 5, 6, and 10 days respectively, giving you an MTTR of five days for that window. At the end of the reporting period, three other incidents remain unresolved. They are not part of the resolved population, so they have no effect on the five-day average, and the headline number looks healthy while your oldest and riskiest exposures sit outside of it.
The same logic applies to code findings, which is one more reason to calculate them separately from incidents.
Why a falling average can hide unresolved risk
A falling MTTR can hide risk because the average only counts items that have been closed, so the longer a critical issue stays open, the longer it stays out of the metric. A team can close a large number of small issues quickly and see the average drop while its most critical findings remain open, and the dashboard will show improvement even though the underlying exposure has not changed.
To keep the metric meaningful, report it alongside measures that capture what it leaves out:
- Open item age: Track how long currently unresolved incidents and findings have been open, with particular attention to the oldest ones.
- Overdue critical findings: Count the high-severity items that are past their remediation deadline and show how long they have been overdue.
- Recurrence: Measure how often the same issue or root cause comes back, which is what the root cause follow-up required by Article 17 is meant to prevent.
- Recovery objective breaches: Record how often actual restoration took longer than the RTO you set for a function.
Together, these give a more complete picture: MTTR describes the work you have finished, while open item age and overdue critical findings describe the risk you are still carrying.
Build evidence around the metric
The evidence that matters most for DORA is the trail behind the number, because Article 17 expects incidents to be detected, recorded, classified, and followed up so that root causes are documented and addressed. Ideally, that trail builds up as part of everyday work, so it is already in place when an examination comes around. We recommend building these elements into your incident and remediation workflow:
- Traceable timestamps for the start and stop events behind every resolution time.
- Clear ownership for each incident and finding.
- Incident and finding records that persist independently of any dashboard.
- Documented remediation actions, together with the verification that confirmed them.
- Exceptions and review decisions, including any accepted risk and the name of whoever approved it.
A well-kept evidence trail is the most convincing way to show that your response and recovery controls operate continuously, and it gives your MTTR the context an auditor will look for.
How does SonarQube help with DORA compliance?
SonarQube fits into the part of the picture that starts in code, where it helps you find and fix issues before they reach production and records each code finding from detection to resolution. SonarQube is the independent verification layer for all code, including AI-generated code, and its deterministic and agentic analysis detects reliability and security issues as code moves through pull requests and CI/CD pipelines. For teams connecting secure development to the follow-up expectations in Article 17, that record shows what was found along with how and when it was resolved.
On the measurement side, SonarQube reports a mean time to resolve for code findings in its project dashboards, which are available in specific editions (the documentation has the details).
The metric shows the average time to resolve issues introduced and detected in the last 30 days and closed within the same period, and it excludes open, accepted, and false-positive issues. In other words, it is a closed-item average with the limitation described above, so we recommend reading it next to the age of your open issues. If you use SonarQube Advanced Security, you can also track MTTR for dependency risks, which covers risks that were fixed or marked as safe during the last 30 days, so it reflects how quickly dependency risks were closed out, which is a broader measure than completed remediation.
SonarQube also supports DORA’s expectations for identifying vulnerabilities (Article 8), controlling software changes (Article 9), and reviewing source code as part of resilience testing (Article 25), which makes it one defensible input to your DORA evidence alongside the rest of your ICT risk management framework.
To see how code remediation reporting works in alignment with your own standards, book a product demo.
Next steps
- How SonarQube enables DORA compliance for financial institutions – a companion read on building security and compliance evidence into the development workflow.
- SonarQube compliance and reporting – how centralized reporting supports obligations in regulated industries.
- SonarQube metric definitions – the exact definitions behind MTTR and related code-finding metrics.
