Home
You send your log and metric data to Tau in JSON format, write rules to find events of interest and act on them.
Acting on an event can be sending an alert (email, webhook etc.) or running code/script (jruby, javascript etc.) or aggregating something in the event (like count unique hostnames over a window of time).
To be more precise, a Rule has condition/s which are evaluated continuously on the Events you send to Tau, if they match the engine can email you, run your code, aggregate something etc.
Tau allows you to all of this and more by just writing rules, no need to worry about writing complex real-time distributed systems.
If you deal with Event data i.e. logs and metrics; the sources can range from IoT (Internet of Things) devices to Application Servers to Security logs, you can use Tau as a platform.
It will solve your problems of alerting and automation at scale. It can be the simple platform that allows you to solve a variety of use cases:
- Application servers are running at 90% CPU for more than 10 minutes, run the auto-scaling command
- If log contains NullPointerException then send an email to xyz@xyz.com with the stacktrace
- Count the number of unique client IPs per webserver for application X; if the count is greater than 500 then send an email and run a Load Balancer throttling script
- If vehicle speed > 65 MPH then log the event to the database (timestamp, user, speed etc logged)
- If 10 washers in 1 location have 10 payment failures in 10 minutes then send an email to the dev team and call technician
It's a
- High performance
- Distributed
- Fault tolerant
- Real-time (streaming)
- Rule based
- Event processing system
It's aimed to simplify the complexity of stream processing, more specifically event correlation at scale.
This diagram shows the high-level components of Tau's engine and how data can flow between them
This is what it looks like with all the nuts and bolts.