Monitoring uptime is very important; and when using microservices it becomes even more important to ensure all services are up and running; and get alerts when they go down. There is nothing worse then your website going down and you hearing about it from someone else. When it comes to downtime, every second counts. Pingdom is one such service which sends alerts when services become unavailable. IngressMonitorController from Stakater automates the management of URL endpoint monitors in Pingdom using kubernetes events.
- How to monitor microservices deployed on kubernetes/openshift?
- How to have automatic monitoring of kubernetes cluster with pingdom?
- How to ensure that web applications deployed on kubernetes cluster are up and available right now using pingdom?
- How to ensure that when a new web application is deployed its monitored automatically using pingdom?
- How to get notified when a web application, REST API or microservice running on kubernetes become unhealthy or unreachable?
- How to have web URL availability monitoring for applications deployed in Kubernetes using pingdom?
- How to have uptime monitoring for web applications or microservices deployed on kubernetes?
- How to have web application availability monitoring when deployed to kubernetes?
- If something deployed on kubernetes does go down, how to know about it immediately?
- How to get notified by SMS or email if your website is down?
- How to proactively monitor your websites and online services & receive immediate notification when a problem is detected so you can quickly resolve the root cause and prevent potential escalation when deployed on kubernetes?.
IngressMonitorController (IMC) offers exactly these features:
- Automatically create new URL endpoint monitors in Pingdom when new applications / services / microservices / REST API's are deployed in Kubernetes
- Automatically update URL endpoint monitors in Pingdom when applications / services / microservices / REST API's are changed/updated in Kubernetes
- Automatically delete URL endpoint monitors in Pingdom when applications / services / microservices / REST API's are deleted/removed in Kubernetes