We've been using ElasticSearch + Kibana for collecting and searching our logs since the beginning of our product. We've recently decided to reevaluate the other monitoring and observability tools we are using and decided to check what Elastic has to offer in this field.
While researching this I found the Elastic solution about Synthetic monitoring is hosting a Heartbeat process somewhere in our infrastructure that generates synthetic calls to the system in order to monitor it. This solves quite well the use case of validating if the system processes are working correctly, but that's only part of the story.
Unfortunately, this solution can't detect networking issues from a specific part of the world or DNS problems from a specific region. For globally used platforms like ours we need to make sure the platform is accessible from each part of the world it is used, and hosting heartbeats machines all around the world ourselves is not really feasible.
Most platforms with mature synthetic monitoring solutions allow you to execute your monitors from their infrastructure which is distributed around the globe, allowing you to not only detect if your application has issues, but also eventual networking problems from specific regions.
Do you know if Elastic have something like this or if we need it, we need to subscribe for a separate service?
Thanks for your message. I'm the product manager for Synthetic and Real User Monitoring here at Elastic. I'm glad to see you're looking into what we can offer! In terms of multiple geo-locations, Heartbeat can be installed and run from multiple geo locations that send data back to your main Elastic Cluster. As Heartbeat is so lightweight, all you need is a micro instance from any cloud provider and you'll be set. There's some fields in the Heartbeat YML for you to name the location that it's being run from and the Uptime interface will show you the Monitor results from all of the available geos.
In terms of a hosted synthetic monitoring service, where we provide the testing nodes and agents, this is a new capability we're adding as we speak. I can't communicate any public release dates but hosted heartbeat pings is an extremely high priority for us and is under active development.
In the meantime, if it is possible to set up those mice instances to run Heartbeat from globally, you'll find it a fantastic solution with resource-based pricing vs per test run pricing.
I hope that clears things up. Let me know if you have any more questions and I'll be happy to help out.
We won't be implementing the solution you proposed mainly because this involves new infrastructure we need to provision and maintain and it's not worth it as there are cheap enough managed solutions we could leverage without so much effort. That said, this is not a super huge problem for us - I was just wondering if I've been missing something.
It is very good news you are working on providing such a managed service as well. I'm looking forward to seeing it rolled out .
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.
.