ClearBreach
On this page

← Playbooks

Your SituationWe had a privacy incident →

In the middle of a breach?

ClearBreach shows whether you must report, to whom and by when, under each law that applies, and writes the reports.

What to do now →
PIPEDAAB PIPABC PIPAAll sectors

Website or Database Compromise, What Canadian Organizations Must Do

By Yong Du · Updated June 14, 2026

Share on LinkedIn

A website or database compromise triggers PIPEDA and PIPA obligations from the moment you discover it, not after the investigation concludes.

⚡ In an active breach right now?

Use the quick reference guide, built for use during an incident.

Open response guide →

This playbook is educational and does not constitute legal advice. It is grounded in the text of PIPEDA, Alberta PIPA, and BC PIPA and published guidance from the OPC, OIPC Alberta, and OIPC BC. If your situation involves regulatory investigation, litigation risk, or circumstances not addressed here, engage a qualified privacy lawyer.

When an attacker gets into your website or database, assume they were there longer, and saw more, than you know. Keep the evidence before you restore, find the first date of entry, change every password stored on the server, and assess what the database held. A dump of personal data posted online almost always has to be reported.

Which laws apply

Jurisdiction Applies when Regulator
PIPEDA The compromised database held personal information used in commercial activity; or the affected individuals are in provinces without substantially similar legislation (Ontario, Manitoba, New Brunswick, and others) Office of the Privacy Commissioner of Canada (OPC), priv.gc.ca
Alberta PIPA The breach involves your organization's activity in Alberta, including its employees there OIPC Alberta, oipc.ab.ca / breachnotice@oipc.ab.ca
BC PIPA The breach involves your organization's activity in BC, including its employees there OIPC BC, oipc.bc.ca

More than one of these can apply. If the threshold is met, report separately to the OPC and the OIPC Alberta where each applies; a report to the OIPC BC is voluntary.


What makes this scenario different

Ransomware, phishing, and lost devices are breach scenarios where the initial event is often obvious, a ransom note appears, an employee reports a suspicious email, a laptop is reported missing. A website or database compromise is frequently the opposite: the attacker gains access, extracts what they want, and leaves no visible sign. Organizations often find out weeks or months after the attacker got in. The news comes from a threat intelligence alert, a customer complaining of fraud, a security scan, or stolen passwords turning up for sale online.

Three features define this scenario.

The attacker may have been inside for a long time, and you may not know how long. Between the first exploit and your discovery, they may have had steady or repeated access to your database. A SQL injection vulnerability in a web form may have been exploited months before discovery. A web shell installed in a compromised plugin may have been active for a year. Your RROSH assessment must account for the full dwell period, not just the moment of discovery.

Exfiltration is often difficult to confirm. SQL queries that extract data look similar to legitimate queries in many logging configurations. A web shell gives an attacker interactive server access with no automatic exfiltration alert. Organizations frequently cannot confirm whether specific records were extracted, only that an attacker had the capability and access to extract them. When exfiltration cannot be ruled out and the attacker had access to personal information, the RROSH assessment should treat exfiltration as likely.

The vulnerability may affect more than you know. A SQL injection vulnerability in one form or endpoint may give access to your entire database, not just the table the form connects to. A compromised plugin on a WordPress site may give an attacker access to every table, every file, and every credential stored on the server. Scoping the breach accurately requires understanding not just where the attacker entered but what they could have reached from that entry point.


Immediate steps, this scenario specifically

1. Take the affected system offline if it is actively compromised, but preserve logs first. If an attacker is actively in your system, taking it offline stops the intrusion. But before making changes, capture the current state: preserve web server logs, database query logs, access logs, and any file system evidence. These are the records your forensic investigation and RROSH assessment depend on. If you patch immediately without preserving logs, you lose the evidence you need to determine what was accessed.

2. Identify the vulnerability and the initial entry point. Determine how the attacker gained access: SQL injection through a web form, a remote code execution vulnerability in a plugin, a compromised admin credential, an unpatched CMS or framework version. The entry point determines the scope of what the attacker could have reached from that position.

3. Determine the dwell period. Check your web server logs, database logs, and security event logs for the earliest evidence of exploitation. If the logs do not show the first entry, use the earliest plausible date. Base it on when the weakness was made public for the version you ran, or on your plugin installation and update history. The dwell period runs from that date to the date you contained the intrusion.

4. Determine what personal information was accessible during the dwell period. Map the database tables and files the attacker could have reached from the entry point. Which tables contain personal information? What categories, customer contact data, payment records, account credentials (hashed or plaintext), health information, transaction history? How many individuals' records were in those tables?

5. Check whether the database credentials or admin passwords were exposed. If the attacker accessed your database or server configuration files, they may have obtained credentials that allow ongoing access, to the database directly, to your email, or to third-party services. Rotate all credentials immediately: database passwords, admin panel passwords, API keys stored in configuration files, and any other credentials accessible from the compromised system.

6. Do not restore from backup before completing your evidence capture. Restoring a clean backup is the right recovery step, but only after you have preserved logs, captured the compromised state, and completed your initial forensic review. Restoring early destroys evidence of what the attacker accessed and for how long.


What drives RROSH in this scenario

Factors that push toward RROSH:

  • The attacker extracted a database dump containing customer or employee records, confirmed exfiltration with malicious intent is strong evidence of RROSH
  • The compromised database contained sensitive categories: payment card data, SINs, health information, account credentials (even hashed, rainbow table attacks are realistic for weak passwords), or enabling combinations (name + date of birth + address)
  • A web shell was present and active during a period when the database held personal information, the attacker had full capability access regardless of whether extraction is confirmed
  • The vulnerability was in a public-facing form or endpoint that handles personal information (checkout form, login page, customer account portal)
  • The attacker's subsequent activity matches monetization of data: credential stuffing against other sites, phishing emails using your customer data, fraud reports from affected individuals
  • The dwell period was extended, weeks or months of potential access

Factors that push against RROSH:

  • The compromised system held only public-facing content with no personal information stored in the database (a brochure website with no user accounts, no forms that collect personal data, no transaction history)
  • The attack was detected and contained within minutes before any database queries were executed, and logs confirm no data-table access
  • The database held only non-sensitive, non-identifying information (aggregated statistics, public content)
  • The attacker's activity was limited to defacement of public pages with no evidence of database access

Once the data is confirmed stolen: when a copy of the database turns up on a criminal forum, or a threat intelligence service confirms it, the test is almost certainly met. The data is in criminal hands and likely being used or sold. Do not delay notification while attempting to negotiate with the attacker or verify the authenticity of the dump. Assess and notify immediately.


Likely verdict range

Scenario Typical verdict
SQL injection confirmed; database dump of customer records exfiltrated RROSH, confirmed access with malicious intent and exfiltration
Web shell present; database with customer personal information accessible during dwell period; exfiltration not confirmed RROSH likely, attacker had full capability access; treat exfiltration as probable
Plugin vulnerability exploited; attacker accessed admin panel but no database queries in logs Case-by-case, depends on what the admin panel could reach and log completeness
Database contained hashed passwords and email addresses only; no financial or health data Case-by-case, depends on hashing strength and whether email + account enables downstream harm
Website defaced; no database access in server logs; no user accounts or personal data in database Likely no reporting obligation, but confirm log completeness before concluding
Payment card data accessed during checkout process (card-skimming script / Magecart attack) RROSH, financial data with confirmed attacker access; also triggers PCI-DSS obligations

Does your compromised site or database need reporting? It turns on what it held and for how long. That is the decision ClearBreach makes, under PIPEDA, Alberta PIPA and BC PIPA at once. See it made on a real case →


Scenario-specific obligations and complications

If RROSH is determined:

  • OPC notification (PIPEDA): Report to the Office of the Privacy Commissioner of Canada as soon as feasible after determining RROSH. Use the OPC breach report form at priv.gc.ca. If your investigation is ongoing, file on available information and supplement the report as your investigation continues.
  • OIPC Alberta notification (AB PIPA): Notify the OIPC Alberta without unreasonable delay. Email breachnotice@oipc.ab.ca. Use the official OIPC Alberta notification form.
  • OIPC BC (BC PIPA): No legal duty to report; a voluntary report is encouraged for a significant breach in your BC activity.
  • Individual notification: Notify each affected individual directly (required under PIPEDA; in Alberta when the commissioner directs, and best sent with the report; expected in BC). The notice should describe what happened, what information was involved, what steps the organization is taking, and what individuals can do to protect themselves. For credential exposure, advise individuals to change their password on your site and on any other site where they used the same password. For payment card exposure, advise them to contact their card issuer immediately.

If RROSH is not determined:

  • No regulator notification required under PIPEDA or Alberta PIPA.
  • No individual notification required.
  • Internal breach record, required under PIPEDA and good practice otherwise: document the intrusion, the scope of access, your RROSH assessment and reasoning, and the remediation steps taken.
  • Investigate the root cause and close the vulnerability. The absence of RROSH does not mean the incident was not serious. It means the specific personal information at risk did not cross the harm threshold.

BC PIPA, voluntary reporting option. BC PIPA has no breach reporting or notification duty, so any report to the OIPC BC is voluntary. For a significant breach in your BC activity, the BC Commissioner encourages a report and expects the people affected to be told where there is a risk of harm. See BC PIPA Breach Reporting.

For Ontario organizations. Ontario has no provincial private-sector privacy legislation, PIPEDA is the applicable framework. Report to the OPC at priv.gc.ca. No separate provincial regulator report is required. See Ontario Data Breach Reporting Requirements.

Complications specific to this scenario:

You do not know when the intrusion started. If logs are incomplete, use the most cautious estimate. That is the earlier of the date the weakness arrived (when the plugin or software version was installed) and the date it was made public for your version. Regulators expect you to use the broadest reasonable window, not the narrowest.

The attacker is threatening to publish or has already published the data. If an attacker is extorting you with a threat to release stolen data, paying the ransom does not satisfy your regulatory obligations and does not guarantee the data will not be released. Assess RROSH based on the fact that the data is in a malicious actor's possession, not based on whether they follow through on their threat. Notify regulators and affected individuals regardless of whether you pay. Involve legal counsel before making any payment.

A card-skimming script was injected into your checkout page (Magecart attack). This is a distinct sub-scenario: rather than extracting a database dump, the attacker injects JavaScript that captures payment card data in real time as customers enter it. RROSH is almost certainly present, payment card data was captured by a malicious actor for every customer who checked out during the skimmer's active period. This also triggers PCI-DSS obligations, which are separate from PIPEDA and PIPA. Contact your payment processor immediately.

Your site runs on a shared hosting environment and the host was breached. If your hosting provider was compromised and your site's data was affected as a result, this is a vendor breach scenario in addition to a web compromise. You remain the accountable party under PIPEDA and PIPA. Assess RROSH based on what your site's database contained and notify accordingly. Your contractual relationship with the hosting provider is a separate matter. See Vendor or Third-Party Breach: What Canadian Organizations Must Do.

Your database credentials appeared in the attacker's dump but you cannot confirm what was accessed. If an attacker posts your database credentials (username/password) in a dump, they had the means to access your database whether or not they ran specific queries. Treat this as potential access to every table those credentials could reach. Rotate all affected credentials immediately and assess RROSH on the basis of what the database contained.


Documents you will need

Regardless of RROSH determination:

  • Internal breach record: date discovered, estimated intrusion start date, vulnerability exploited, scope of database access, personal information categories affected, number of individuals, RROSH assessment and outcome, remediation steps
  • Web server logs and database query logs for the dwell period (preserved before remediation)
  • Forensic summary from your investigation (internal or third-party)

If RROSH is determined, also:

  • OPC breach report (PIPEDA)
  • OIPC Alberta breach notification (AB PIPA), if the breach involves your Alberta activity; email breachnotice@oipc.ab.ca
  • OIPC BC voluntary report (BC PIPA), if the breach involves your BC activity and is significant
  • Individual notification letters, one per affected individual or one per household

ClearBreach tells you which of these you need, and gives you the regulator reports and notification letters from your assessment answers.


Common mistakes in this scenario

Restoring from backup before saving the evidence. Wanting to get back online fast is natural. But restoring before you capture the logs and the compromised system destroys the evidence your assessment depends on. Preserve first, restore second.

Waiting for forensics to complete before notifying. A full forensic investigation of a web compromise can take weeks. Regulators expect notification as soon as RROSH is determined, not when forensics closes. File your breach report on available information and update as your investigation continues. Supplementing a timely report is always better than a complete report filed months late.

Scoping the breach based only on confirmed queries in logs. If logs show an attacker executed specific queries, those queries define a confirmed minimum scope. But incomplete logging, log gaps, or log tampering by the attacker means confirmed queries may understate actual access. Scope your RROSH assessment based on what the attacker could have accessed from the entry point, not only what your logs can confirm.

Assuming a hashed password database does not trigger RROSH. Hashed passwords are not plaintext passwords, but they are not safe either. MD5 and SHA-1 hashed passwords are crackable. Even bcrypt-hashed passwords are at risk if passwords are weak. A database dump of email addresses and hashed passwords in an attacker's hands enables credential stuffing attacks against every other site where those individuals use the same email and password. This is real harm, not theoretical harm.

Treating a plugin update as the end of the incident. Patching the exploited vulnerability closes the door, but it does not address what happened while the door was open. Update the plugin and then complete your RROSH assessment for the dwell period. Both are required.

Not changing passwords after a server compromise. An attacker on your server may have every credential stored there: database passwords in configuration files, API keys, admin passwords. Change them all. Failing to rotate all of these creates ongoing risk even after remediation.


What regulators have found

OIPC Alberta: A national retail chain's online checkout page was compromised by malicious code that silently captured payment card data and personal information from customers completing transactions. The code was active for 60 days before detection, from November 19, 2022 to January 18, 2023. Card numbers, expiry dates and security codes of 1,877 Albertans who bought during that window were treated as compromised. OIPC Alberta required notification to all affected individuals under PIPA. The 60-day window is the defining feature: because malicious checkout code does not cause visible disruption, there is no operational signal to trigger investigation. Detection requires active monitoring. It does not happen through normal use of the site.

OPC (federal): Canada's Privacy Commissioner investigated MGM Resorts after a 2019 breach of guest data, including government ID numbers for 5,635 Canadians, which was later posted for sale on a dark web forum. MGM did not begin assessing the impact on Canadians until about eleven months after learning of the breach. The OPC found it had failed to assess promptly whether there was a real risk of significant harm, and so had not reported or notified as soon as feasible. The lesson for any database compromise: the duty to assess starts when you learn of the breach. Waiting for a complete forensic picture before assessing is itself the failure.

Sources: OIPC Alberta, Decision P2023-ND-004, Breach notification decisions; OPC, PIPEDA Findings #2022-004, Investigations into businesses


MSP note

If you manage web infrastructure or databases for clients and this compromise occurred in a client's environment you administer:

Tell the client at once, with the technical detail: which weakness was used, what the attacker reached, how long they seem to have been in, and what you have done or are doing to fix it. The client organization is the accountable party under PIPEDA and PIPA for their individuals' personal information.

If the compromise resulted from a vulnerability in software you deployed and managed, your contractual liability to the client is a separate matter from the client's regulatory obligations. Both run concurrently.

If your own systems were the way in, such as a shared platform you run for several clients, you may be the vendor in several breaches at once, one for each client affected. Each affected client organization needs to be notified separately and will conduct their own RROSH assessment. Run a ClearBreach assessment for each affected client under your MSP account and document each separately.


Ready to assess this breach? Find out whether you must report it, to whom and by when, under each law that applies, with the regulator reports and notification letters to send. See it done on a worked case →

The demonstration is the real assessment, already answered for a fictional company, with three answers left for you to change. It ends on the verdict and the documents it produced. See a real verdict → to run it on your own breach.


This playbook covers PIPEDA, Alberta PIPA, and BC PIPA obligations for private-sector organizations. If your organization handles personal health information under provincial health legislation, such as Alberta's Health Information Act or BC's E-Health (Personal Health Information Access and Protection of Privacy) Act, additional obligations may apply when a database containing health information is compromised. Those obligations are not covered here. A Magecart or card-skimming attack also triggers PCI-DSS obligations, which are separate from PIPEDA and PIPA and are not covered here.


Frequently asked questions

Our website was hacked and customer data may have been accessed, what are our obligations?

Your first obligation is to contain the intrusion and preserve evidence. Your obligations under PIPEDA and Alberta PIPA start running when you determine a breach has occurred, which is usually soon after discovery, not when your forensic investigation is complete. You must assess whether the breach poses a real risk of significant harm (RROSH) to affected individuals and, if so, notify the applicable regulator and affected individuals as soon as feasible. You do not need to wait for a full forensics report before notifying, regulators expect notification when RROSH is determined, not when the investigation closes.

Attackers used SQL injection to access our database, does that change our obligations?

The attack method does not change your obligations, it informs your RROSH assessment. SQL injection that successfully extracted records from a customer or employee database is a confirmed access event with data exfiltrated to an attacker. Confirmed exfiltration by a malicious actor is one of the strongest indicators that RROSH is present. The attack being technically sophisticated does not reduce your reporting obligations; if anything, the targeted, intentional nature of the access is a factor that pushes toward RROSH.

We found a web shell on our server, does that mean data was accessed?

A web shell is a backdoor that gives an attacker persistent, interactive access to your server. Its presence confirms unauthorized access. Whether data was specifically extracted depends on what logs show, but the presence of a web shell means an attacker had the capability to access every database and file the web server could reach, for the entire period the shell was active. If that period includes your customer or employee database, treat the data as potentially accessed and assess RROSH on that basis. The absence of confirmed exfiltration does not mean data was not taken.

We patched the vulnerability, do we still need to report?

Yes. Patching the vulnerability stops the active intrusion but does not extinguish your regulatory obligations for the breach that already occurred. Your PIPEDA and PIPA assessment covers the period from when the vulnerability was first exploited to when it was remediated. If RROSH is present during that window, you notify. Patching is a required remediation step, it is the beginning of your response, not the end of it.

Ready to assess this breach?

See whether you must report, to whom and by when, under each privacy law that applies to you, with the reports and letters to show for it.

How ClearBreach handles a breach →

See plans and pricing

Privacy case of the month

Once a month, a five-minute read: a real case a Canadian privacy commissioner investigated, what went wrong, and what to check in your own organization so it doesn't happen to you. Written for organizations under federal, Alberta and BC privacy law. Not written for Quebec.

We send one email to confirm. Nothing else arrives until you confirm, and an unconfirmed address is deleted after 7 days. Sent by ClearBreach Technologies Inc. Mailing address and contact details · Privacy policy