Updated The Telegraph newspaper managed to leak 10TB of subscriber data and server logs after leaving an Elasticsearch cluster unsecured for most of September, according to the researcher who found it online.
The blunder was uncovered by well-known security researcher Bob Diachenko, who said that the cluster had been freely accessible "without a password or any other authentication required to access it."
After sampling the database to determine its owner, Diachenko saw the personal details of at least 1,200 Telegraph subscribers along with a substantial quantity of internal server logs, he told The Register.
"A significant portion of the records were unencrypted," he said. Screenshots he provided showed information including the user-agent string and device type, while categories of personal data included subscribers' first and last names, email addresses, subscriber status, IP addresses and device type and operating system.
Affected users "should be on the lookout for targeted phishing and scams," Diachenko advised. "Names and emails in the database can be used to send readers targeted scam messages."
Aside from potential scam emails, the risk from this breach is relatively low unless having your news-reading habits collated in one place might cause professional embarrassment: Diachenko highlighted that in the data sample he viewed were a handful of gov.uk email addresses.
We have asked the Telegraph whether it has notified affected subscribers and will update this article if the media organisation responds.
Diachenko said he found the Elasticsearch cluster on 14 September, eventually getting the attention of the Telegraph which closed off public access two days later. He theorised that it had been exposed online for most of September.
- Indonesian authorities probe million-record leak from national COVID app
- Dowden out, Dorries in: Is UK data protection in safe hands?
- Database containing personal info on 106m people who traveled to Thailand found open to the internet – report
- Amazon Elasticsearch Service is so flexible it wants to be called by a new name
Unsecured Elasticsearch clusters are relatively common ways for personal or sensitive data to be exposed to the wider world. At the same time as Diachenko reckoned the Telegraph's cluster became publicly accessible, Indonesian authorities were investigating a million leaked records that originated from an Elasticsearch installation powering its eHAC vaccine passport app. Microsoft managed to do something similar with 6.5TB of Bing search data a year ago.
Similarly, Diachenko himself – an expert in finding unsecured databases online – spotted 106 million travellers' details freely accessible online through a data store in Thailand. That contained 10 years of travel history relating to people entering and leaving the nation.
With the rise in the 2010s of dedicated search engines such as Censys and Grayhat Warfare, it has become ever easier for researchers to find data stores that aren't properly secured and draw their owners' attention to them. Unfortunately, that also means nefarious people can use the same tools to extract their contents.
Elasticsearch itself made a foray into the security market a couple of years ago, launching its own SIEM product after seeing others building similar suites on top of its database technology. ®
Updated at 14:30 UTC, 5 October 2021 to add:
A Telegraph Media Group spokesperson told The Register: "We are aware that there was a time limited exposure of technical data, the vast majority of which had no meaning. Within that data was a small amount of system login data - less than 0.1 per cent of our subscriber/registrant database.
"We take this matter extremely seriously and took quick action to resolve this issue and close down the exposure. We have been in touch with the small number of those who have been affected to update them."