How does HubSpot make its system reliable and resilient?
We build our software platform so that HubSpot is available and accessible in a variety of disaster scenarios. Each service in our micro-services architecture is spread across multiple servers running in different data centers. Our services use API's to communicate with each other to reduce the interdependence between each other. Every service also has a corresponding test environment where changes are deployed before they are migrated to production.
HubSpot also utilizes a worldwide Content Delivery Network (CDN) to distribute content to a location closest to users enabling quick and consistent access wherever you are.
Are HubSpot’s site and services always available?
Our goal is that you can always access your HubSpot account. There are times when the HubSpot service will be unavailable due to planned maintenance or due to a component failure. In such cases, HubSpot staff are paged as soon as the failure is detected and work to make sure the service is back up in the shortest possible time. You can check our Status site for the latest status of our services.
How does HubSpot make sure outages due to component failures do not reoccur?
When an outage or significant failure occurs, HubSpot's primary goal is to get the service up and available to customers. After the issue has been resolved, the team that owns the affected service holds a post mortem, i.e. a formal review of the incident. During this review, teams use “The 5 Why’s” process to analyze the root cause of the event and to develop a list of immediate action items to make sure this event, and other events like this one, do not re-occur. Action items are expected to be completed within 30 days of the event. Each post mortem event is documented in detail and future learnings are incorporated into long term plans.