On June 21, 2022, the US-based global content delivery network (CDN) provider and security company Cloudflare suffered an outage at 6:27 UTC that lasted until 7:42. The outage was caused by a network configuration error that affected 19 of Cloudflare's data center locations — Amsterdam, Atlanta, Ashburn, Chicago, Frankfurt, London, Los Angeles, Madrid, Manchester, Miami, Milan, Mumbai, Newark, Osaka, São Paulo, San Jose, Singapore, Sydney, Tokyo.
In the early days of log analysis, application developers would use their logging libraries to write logs to files stored on a disk. After years of relying on those libraries, they found that they were unable to monitor the performance of their applications anymore because they didn’t understand the way their logging libraries worked. This led to a shift from using log files stored on a disk to using Syslog.
For the vast majority of organisations, some form of round-the-clock cover is critical to successful business operations. On-call is an essential part of an effective incident response process, yet there is no commonly accepted playbook on how to most effectively structure and compensate on-callers. We ran a survey to uncover the mysteries of how on-call works in organisations of different shapes and sizes around the world.
Honeycomb uses AWS Lambda as a core part of our query execution architecture; Lambda’s ability to quickly allocate lots of resources and charge us only for use is invaluable to keeping Honeycomb fast and affordable. Our total Lambda bill is easily accessible in the AWS Console, but how do we know which customers or application areas dominate this bill? How do we judge the cost of changes we make to our own software?
Monitoring performance in front-end applications is vital. It focuses on the areas of the application users experience. These areas are slow rendering or frame rate, network request errors, and unresponsive user experience. Your application’s users expect the experience of using the application to be fast, responsive, and smooth. In the first article of this series, we discussed some tips for optimizing your React Native application performance.
In a world where the customer’s digital experience is critical to business outcomes, it is crucial to understand how our applications are behaving. As businesses increasingly rely on the performance and availability of revenue-generating applications, the tolerance for downtime and slow response times has plummeted – so the response to production issues must be quick and effective.
So here we are…the final chapter. In Part 2 of this series, we started to drill down into some of the concepts that make Kafka great. We concluded that although terminology between MQTT and Kafka was similar (for example topics), they behaved quite differently under the hood. We also took a brief overview of Kafka Connect and how we can use some of the enterprise connectors to stream our data to other platforms. Yet we did learn that Kafka does have some shortfalls.
Morph is a new database migration tool designed to improve Mattermost schema migrations. It can be used programmatically with a library and from a CLI. Morph stores the schema version in a table, where all applied migrations are persisted. It’s a flexible open source library that can be used with a driver interface implementation and a source to read migration scripts from. The engine uses a dependency injection pattern so that any implementation can be used with the library.
