Option to store results of previous HC runs
wimpers opened this issue · 1 comments
wimpers commented
Now the HC is a view at that time of the state of the cluster. It would make sense to keep track of the results of previous runs of the HC. Many options are possible, store locally or in Arakoon. This way we could for example raise an alert only in case the DTL is degraded for longer than X minutes.
Items where we could use this:
- DTL degraded
- MDS safety
- disk safety
JeffreyDevloo commented
Can be done when using the celery scheduling and letting the healthcheck auto-refresh its result within a certain time period.
Also this kind of reminds me of #385