ISO New England Inc.’s Post

In light of the recent CrowdStrike outage, it's a perfect reminder of the immense dedication and expertise that SysAdmins, Site Reliability Engineers (SREs), and Cloud Engineers bring to our organizations every day. When systems go down, these professionals work tirelessly behind the scenes to restore services, often going above and beyond to ensure minimal disruption. Their ability to troubleshoot, problem-solve, and implement effective solutions in high-pressure situations is nothing short of remarkable. Here is a great article explaining how to fix Crowdstrike BSoD on Azure-hosted VMs https://lnkd.in/ed5ccub3 #SysAdmin #SRE #CloudEngineers #TechHeroes #ITprofessionals #CrowdStrike #Outage #Resilience #Innovation

See how the US Navy deployed a DevSecOps environment in AWS Secret Region to deliver new capabilities for sailors! 🚢 This post from the @AWSCloud Blog has all the details! Contact Brick Bridge Consulting to learn more.

This. Usually at a company, there are more employees than security team members. For the forces of good to have a chance, security must be deeply rooted in your company, and a priority. You can have all the tech and all the budget you want, but if it’s your security team against everyone and everything with no allies, it all becomes so much harder. You can change that by having a top-down culture of security. More about how we do it at AWS (and other interesting tidbits) below! https://lnkd.in/ghyyxnCY

Are you prepared for the next outage? The CrowdStrike outage is a wakeup call for everyone in a leadership position (again). We're all wondering what happened particularly when there are some are some obvious solutions to the *technical* problem. 1. Testing Lab with Automated CI/CD Test every configuration of common customer configurations for each release when you submit a MR/PR or something merges. I've done this before and it's expensive but worth it at scale. 2. Incremental rollout. Just don't deploy to everyone at once. Thundering Stampede and all of that. Canary Releases/Blue Green are a few techniques. 3. Provide tooling to your developers to actually test realistic environments: This is not just a "throw it at QA" problem... When developers have the ability to _easily_ test and build different configurations the feedback loop is quicker and many issues will be resolved during dev time. --- These kinds of failures go way beyond just a single person committing some code. Systemically it's hard to invest in testing until sh*t like this happens. Hopefully this is a lesson for the industry as a whole (like the Amazon outage in us east way back). What has worked for you? #CyberSecurity #DevOps #CI #CD #AtScale

AWS re:Invent is here! Join Peter Wrenn and I at booth 900 at 4pm today to learn how Tines & our partner Snyk can empower security teams to eliminate repetitive work within the DevSecOps lifecycle. #awsreinvent2023

These moments are great to step away from the vendors in question, and think about the multiple factors that have to coincide in order to cause an outage, and even worse a prolonged/widespread outage. 1) Production systems are updating directly from the internet, onto a current/latest version. This means that your internal systems did not have a chance to take this update, test it against your systems, and, if it were to cause an issue... do so in an isolated and non-impacting environment. 2) Recovery processes being slow or incomplete. The stated repair from CrowdStrike here is to go remove a specific file. Which is fine, for individual systems that can't be wiped. But if you're trying to handle all of your kiosks, servers, etc.. there needs to be a fully tested image restore process. Hell, talk to the universities. Most of them wipe and restore their lab PCs daily. Point being, there are well-known ways to wipe and recover systems in minutes, if not hours. (caveat of course is where is the data?) 3) Lastly is the issue of the vendor. Because, 1 and 2 are under an organization's control. There are plenty of software updates that go out regularly, and as much as Crowdstrike gets blamed today, it is organizational failures of 1 and 2 that caused the outages. If it wasn't Crowdstrike, it'd be some other vendor. When dealing with large scale failures, its almost never the last or most visible thing that caused it. Its a combination of factors, starting with much earlier decisions. #windows #nochangefriday #outage #systemaccidents

Want to flag anomalies within your AWS environment and overlay them with MITRE ATT&CK? We have extended our analytics engine to process CloudTrail logs and highlight 40+ threats. Here's a full list of detection use cases >> https://lnkd.in/e_RShSBn

DAY 56/100 FOR INTERVIEW PREPARATION OF DEVSECOPS QUESTION:- A user is storing objects in Amazon S3. The user needs to restrict access to the objects to meet compliance obligations. What should the user do to meet this requirement? ANSWER:- Tag the objects in the S3 bucket While tagging is useful, the most common and effective method to restrict access to S3 objects for compliance purposes would typically involve using S3 bucket policies, IAM policies, and potentially enabling S3 Object Lock for compliance with data retention requirements. Therefore, none of the provided options perfectly aligns with best practices for restricting S3 access. The focus should be on policies rather than tags or network controls.

Our very first blog post!! Check out this inaugural blog written by Noah Phillips on our innovative DevSecOps technology - if you are at #WEST2024, better yet come on by GitLab’s booth 1136…you can see BEDDE in person!

Do you use S3? A recent blog shared by Maciej Poćwierz shows that AWS charges you for unauthorized 4xx calls on your buckets too. 🤯 Another gentleman figured out that there are around 65k s3 bucket names visible in GitHub for us-east-1 region alone -- this could be a big bill (+security risk?) if someone tries to DDoS your buckets... Source: https://lnkd.in/gZJ252TT Want to read a discussion around it? Check https://lnkd.in/gAqNBnmc What are some other practices by companies that you've found to be odd? #engineering