Infinispan is a caching and data grid solution — it means that there can be thousands of servers connected into one cluster.
Infinispan uses JGroups protocol for internal communication between servers. It is very hard to spot bugs and other network related problems in huge text logs (Terabytes of plain text). Therefore, our intention is to provide new tool that can be used by users and developers! for spotting issues and regressions immediately while operating and testing with Infinispan cluster.
The objectives of this project are to:
- Provide graphical representation of network communication between Infinispan nodes (but not only Infinispan nodes — we can go for other “targets” in the future)
- Provide mechanism how to spot regression and issues more easily
- Provide historical records of network communication in a graphical way
- Design convenient UI
- Deal with performance issues effectively (monitoring of a big cluster can be very resource demanding)
- Prepare Infinispan and/or JGroups for monitoring from our Web Console
See how it could possibly look like in the future. We can very easily see that Server 15 (S 15 — red color) is the main sender. It sends a lots of data to other nodes. Pipes are going from S15 to other servers. On the other hand Server 16 (S 16 — turquoise color) is the main receiver, it is mainly receiving data from other servers in the cluster.
Your main contact person for this project is Tomáš Sýkora (firstname.lastname@example.org).