Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- you are an sre/devops that owns a web server
- this web server can throw errors, and it self reports error rates
- the error rates are returned at discrete time steps (for example, every 5 minutes)
- you want to know if the server is unhealthy, that way you can stop playing ping pong
- a server is unhealthy if the average error rate in a period exceeds some threshold
- question:
- determine if the server is unhealthy at a particular time
- example:
- error rates [ 0.1 0.2 0.9 0.9 0.3 0.1 0.01 ]
- these are error rates at time 0, 1, etc
- for a period of length 2, and time 1, with threshold 0.5, the server is unhealthy
- for a period of length 3, and time 6, with threshold 0.3, the server is healthy
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement