Inspire by the post by Martin Fowler, wanted to share my thoughts as well.
You need to verify the maturity of you team and infrastructure before jumping into microservices.
Do your have a monitoring system to identify service failures? Which system is broken?
In case of failures you might need to provision new machine or containers. So you need to have infrastructure that enables you to launch new service in a couple of hours.
You need to have CI/CD up and running as you need to be able to deploy each services as fast as possible. Which is you would need to do anyway for service provisioning.
Identify bugs and failures
You need to have log explorer, and request-id propagation to identify bugs of your application. Otherwise it would be a nightmare to trace bugs across systems.
You surely need tests to guard against your system boundaries, so that any change there should not break the system.
Cost effective infrastructure
When you start breaking into microservices, all of a sudden you will find a lot of services which are so small that it is not cost effective to put them in different machines. Eventually you would move into a solution like docker to able to reduce the cost for you jenkins and all your production instances. And also solve the issue of running a bunch of services to develop a feature. On-boarding new developers become an issue as well.
These are some of the issues that you need to consider before jumping into microservices.