DevOps Articles

Content about software operations and DevOps.

Separate Your Monitoring From Infrastructure

Separate your monitoring from your primary infrastructure to avoid dangerous blind spots. Cloud providers' internal monitoring fails when their infrastructure fails, leaving you debugging outages without visibility. TrackJS hosts on dedicated hardware separate from major cloud providers for true ...

Todd H. Gardner
July 30, 2019

[Read more »](/content/blog/separate-monitoring/ "Separate Your Monitoring From Infrastructure"/index.html)

Metered Billing for Monitoring is Anti-Developer

A fictional tale exposing the anti-developer nature of metered logging: Sarah gets locked out of LogCorp during a critical outage because adding debugging logs pushed them over quota, forcing an emergency upgrade during crisis. TrackJS rejects this model, pricing by site traffic instead of error ...

Todd H. Gardner
July 16, 2019

[Read more »](/content/blog/metered-logging-is-anti-developer/ "Metered Billing for Monitoring is Anti-Developer"/index.html)

Moving to Mailgun

TrackJS switched from failing Mandrill to Mailgun after years of increasing email delivery problems. Key lessons: make failures visible through chat alerts, avoid abstractions until you need them, and sometimes big changes become unavoidable when dependencies deteriorate.

Eric Brandes
October 09, 2018

[Read more »](/content/blog/moving-to-mailgun/ "Moving to Mailgun"/index.html)

Learning from GitLab: Making Production Obvious

Learn from GitLab's outage by making production environments visually obvious. TrackJS shares Ansible snippets to turn Linux command prompts and Windows UI chrome bright red in production, preventing dangerous "wrong server" mistakes.

Jordan Griffin
March 27, 2017

[Read more »](/content/blog/learning-from-gitlab-making-production-obvious/ "Learning from GitLab: Making Production Obvious"/index.html)

Outages and Our Way Forward

TrackJS has experienced unacceptable infrastructure problems due to Azure instability. We're migrating to dedicated hardware with automated infrastructure to improve availability and performance in the coming weeks.

Todd H. Gardner
April 29, 2015

[Read more »](/content/blog/outages-and-our-way-forward/ "Outages and Our Way Forward"/index.html)