An MFA fatigue attack, also called push bombing, happens when someone with stolen login credentials floods a victim with authentication prompts until they approve one out of frustration. Phishing-resistant MFA, such as FIDO2 security keys, is the strongest protection available, with number matching as a critical interim step. If you’re reading this after a prompt storm, reset the exposed credentials and investigate immediately.
TL;DR:
- An unexpected prompt usually signals exposed credentials; deny it, reset the password, revoke active sessions, and investigate recent sign ins, email rules, and file activity.
- FIDO2 security keys offer the strongest defense, while number matching is the immediate interim choice; advanced adversary in the middle kits can still steal live sessions.
- For accounts still using push approval, enable number matching and disable one tap approval, then add rate limits and block unfamiliar devices or locations.
- Alert on repeated MFA challenges without a successful login, especially overnight or from unfamiliar locations; denied approvals indicate compromised credentials and warrant immediate investigation.
Table of Contents
- What is an MFA fatigue attack?
- How MFA fatigue attacks work (attack chain and common vectors)
- Why these attacks still succeed: human and organisational factors
- Types of MFA and their relative strengths and weaknesses
- Detecting MFA fatigue: logs, telemetry and correlation rules security teams must enable
- Practical mitigations and controls (prioritised checklist)
- Incident response playbook for an MFA fatigue event
- Practical checklist and TechBug perspective for operationalising defences
- Author perspective: priorities for security leadership
- How we help stop MFA fatigue attacks
- FAQ
- Sources
What is an MFA fatigue attack?
An MFA fatigue attack, sometimes called push bombing or MFA spamming, is a technique in which an attacker who already holds a valid username and password repeatedly triggers authentication requests to the account owner’s device. The goal is simple: wear the person down until they tap “approve” just to make the notifications stop.
This attack cannot happen in a vacuum. It requires a precondition, almost always compromised primary credentials obtained through phishing, credential stuffing or self-service password reset (SSPR) abuse. Without a working password, there’s no login attempt to generate the push in the first place.
Success, from the attacker’s side, looks like:
- A completed login that grants an active session inside the target environment
- Lateral movement into email, file shares or business applications
- Persistence through registered devices or session tokens that outlast the initial breach
How MFA fatigue attacks work (attack chain and common vectors)
The attack chain is methodical, not opportunistic. Understanding each stage helps security teams know where to intervene.
- Credential theft. Attackers harvest usernames and passwords through phishing campaigns, credential stuffing against leaked password lists, or by purchasing dark-web credential sets.
- Identity provider exploitation. With valid credentials, attackers log into the identity provider or SSO portal, which automatically triggers a push notification, or abuse self-service flows to generate repeated prompts.
- The flood. Prompts arrive every few seconds or minutes, sometimes for hours, designed to exhaust the target’s patience rather than their suspicion.
- Social engineering layer. Some campaigns pair the flood with a phone call or message impersonating IT support, asking the target to “just approve the one you’re waiting on.”
- Amplification via AiTM kits. More advanced operators use adversary-in-the-middle proxy phishing kits that sit between the victim and the real login page, capturing session tokens in real time rather than relying purely on prompt fatigue.
Persistence patterns vary: some attackers give up after a handful of failed attempts, others return across multiple days, testing whether the target’s vigilance has dropped.
Why these attacks still succeed: human and organisational factors
Push bombing exploits a predictable human response to repeated interruption. A person getting a dozen prompts at 2am assumes it’s a glitch, not an attack, and taps approve to get back to sleep. Poor prompt design makes this worse: many authenticator apps show no origin details, just a bare approve or deny button, defaulting to one-tap convenience over scrutiny.
Underneath the human factor sit configuration gaps that make the attack possible at all:
- No rate limiting on authentication requests
- Number matching left disabled
- Conditional access policies that don’t account for unfamiliar locations or devices
Pro Tip: Training tells staff what to do; only technical controls stop the attack from reaching them in the first place. Run both together.
Types of MFA and their relative strengths and weaknesses
Not all MFA is equal against push bombing, and knowing the hierarchy matters when you’re prioritising a rollout.
- Phishing-resistant MFA (FIDO2/hardware security keys): requires physical possession and a cryptographic handshake with the legitimate site, which CISA names as the strongest available protection against push bombing and credential phishing alike.
- Authenticator app push with number matching: forces the user to enter a code shown on the login screen, tying approval to an active session rather than a blind tap.
- Authenticator app push without number matching: vulnerable to fatigue attacks because approval requires nothing more than a single tap.
- TOTP and hardware tokens: solid for everyday use, though they rely on the user correctly entering a rotating code rather than preventing the push flood itself.
- SMS and voice codes: the weakest tier, exposed to SIM swap and SS7 interception, and best treated as a last resort rather than a primary method.
Detecting MFA fatigue: logs, telemetry and correlation rules security teams must enable
Early detection depends on correlating signals across identity and network logs rather than waiting for a successful breach to surface elsewhere. The clearest indicator is volume: multiple MFA prompts in a short window with no successful login, especially outside business hours.
Watch for:
- A spike in authentication challenges tied to a single account, particularly overnight or on weekends
- Login attempts originating from an unfamiliar IP address or geolocation inconsistent with the user’s normal pattern
- Rapid retries within seconds of each other, suggesting automated or scripted triggering rather than a forgotten password
- Denied push approvals, which often indicate the attacker holds a valid password but the real user is refusing the prompt
The relevant log sources are your identity provider or SSO platform, authenticator app telemetry and conditional access logs. MITRE ATT&CK’s T1621 technique page documents detection strategies built around alerting on excessive MFA challenges without matching successful logins, and correlating challenge frequency against anomalous origins.
A denied approval is not a false alarm to dismiss. Treat it as a live indicator that credentials are compromised, and begin investigating the source IP, the account’s recent activity and any linked sessions immediately rather than waiting for a second event.
Practical mitigations and controls (prioritised checklist)
Mitigation works best as a sequence: interim fixes now, architectural change over the medium term. Here’s the order that matters.
- Migrate to phishing-resistant MFA where feasible. FIDO2 security keys or platform authenticators eliminate the push-approval step that attackers exploit, and remain the control CISA recommends first.
- Enable number matching immediately for every account still on app-based push. This is the interim measure CISA mandates when phishing-resistant MFA isn’t yet deployed everywhere, and it closes the gap fastest for the least effort.
- Disable one-tap approval in any authenticator app settings that still allow it.
- Apply rate limiting on authentication attempts so an account can’t be flooded with dozens of prompts in minutes.
- Tighten conditional access and risk-based policies, flagging or blocking sign-ins from unfamiliar devices, locations or impossible travel patterns. Our guidance on Microsoft 365 security controls covers how to configure these settings inside Entra ID.
- Maintain IP and device blocklists that update automatically from threat intelligence feeds.
- Give users a one-click reporting channel. Microsoft Authenticator’s report suspicious activity feature lets a user flag an unexpected prompt directly, and when that report is tied to automated remediation, it meaningfully shortens response time.
- Automate the response. A reported or detected fatigue attempt should trigger a forced password reset and session revocation without waiting for manual review.
- Enforce device registration so only known, managed devices can complete authentication.
Pro Tip: If a prompt appears that you didn’t initiate, deny it, change your password immediately and report it. Never approve “just to make it stop.”
Incident response playbook for an MFA fatigue event
Once a fatigue attack is confirmed or strongly suspected, speed matters more than process polish.
- Contain immediately. Reset the affected account’s password, revoke all active tokens and sessions, and temporarily step up to stricter MFA for that user.
- Investigate. Correlate identity provider logs with IP and device telemetry, and pay close attention to any approved pushes, not just denied ones, since an approval means the attacker likely gained a session.
- Check for lateral movement. Review email rules, file-sharing activity and app sign-ins from the compromised account for signs the attacker went further than the initial login.
- Recover. Force re-enrolment of MFA methods, rotate any credentials that may have been reused elsewhere, and apply phishing-resistant protections to the account going forward.
- Harden afterwards. Update conditional access policies, run targeted training for the affected team, and schedule a tabletop exercise so the next event is handled faster.
Practical checklist and TechBug perspective for operationalising defences
Rolling out phishing-resistant MFA across a real organisation works best in phases rather than a single cutover.
- Assess first. Map which accounts carry the highest risk, admins, finance, and anyone with access to sensitive data, using a structured cyber risk assessment.
- Pilot with high-risk users. Deploy security keys or number matching to the smallest group that matters most before a wider rollout.
- Phase the remainder. Expand in waves, checking for help desk friction at each stage.
A 90-day remediation plan typically covers assessment and quick wins (number matching, rate limits) in the first month, pilot deployment of hardware keys in the second, and organisation-wide rollout with monitoring in the third. Our managed detection work includes proactive monitoring, alert triage and incident handover, as outlined in our overview of what managed IT gives SMBs. Call in outside support when the incident involves lateral movement or multiple accounts; smaller, single-account events can often be handled internally with the right logging in place.
Author perspective: priorities for security leadership
Phishing-resistant MFA deserves a line item in next year’s procurement budget, not a someday project. Fix the cheap things first: number matching, rate limits, disabling one-tap approval. Then sequence the architectural shift to hardware keys for your highest-risk accounts. The payoff isn’t dramatic, it’s fewer incidents to chase and faster containment when one slips through.
— Ru
How we help stop MFA fatigue attacks
We approach identity hardening the way CISA and MITRE frame it: phishing-resistant MFA first, number matching and rate limiting as the interim layer, and monitoring that catches the gap in between. We recommend the security key or authenticator setup that fits your existing systems, adapting to different vendor options.

An engagement typically starts with an assessment of your current MFA posture, moves into remediation (number matching, conditional access tuning, key rollout), and settles into managed monitoring so prompt storms get caught and escalated before they become account takeovers. If a credential manager like the one compared in Dashlane vs 1Password is part of your hygiene plan, we can factor that into the broader setup too.
Get in touch through our IT security services page to book an assessment, or see how our approach aligns with the Essential Eight framework.

FAQ
What is the difference between MFA fatigue and MFA bombing?
They’re the same attack described two ways: MFA fatigue describes the psychological effect on the victim, while MFA bombing or push bombing describes the attacker’s technique of flooding approval requests. Both terms refer to the same underlying method of exploiting one-tap push approval.
Does number matching fully stop MFA fatigue attacks?
Number matching significantly reduces the effectiveness of random push floods by requiring the user to enter a code shown on an active login screen, but it isn’t absolute. Advanced AiTM proxy kits can still capture live sessions in some cases, which is why phishing-resistant MFA remains the stronger long-term control.
What should I do if I get an MFA prompt I didn’t request?
Deny the prompt immediately, change the account’s password and report the activity through your organisation’s reporting channel, such as Microsoft Authenticator’s report suspicious activity feature. A single unexpected prompt usually means your credentials are already compromised, so don’t wait for a second one before acting.
How can I tell if my organisation is under an MFA fatigue attack?
Look for a burst of MFA challenges on one account with no successful login, especially from unfamiliar IP addresses, since MITRE’s detection guidance treats this pattern as a core indicator of T1621 activity. Denied approvals are just as significant as approved ones and warrant immediate investigation.
Which type of MFA is most resistant to push bombing?
Phishing-resistant MFA, particularly FIDO2-compliant hardware security keys, is the most resistant option because it requires physical possession and a cryptographic exchange rather than a simple approve or deny tap, a hierarchy CISA lays out explicitly. SMS and voice codes sit at the opposite end, vulnerable to SIM swap and interception.
Sources
- Implementing number matching in MFA applications (CISA fact sheet)
- Multi-Factor Authentication Request Generation (T1621) — MITRE ATT&CK
- How to configure MFA settings — Microsoft docs
- MFA fatigue: What it is and how to respond — University security guidance
- Phishing attacks that defeat MFA are easier than ever — Ars Technica
