Package: gmetad Version: 3.6.0-1 Severity: wishlist Tags: upstream Currently, when gmetad receives a metric, it creates an RRD file for the metric and starts storing all subsequent values received. All RRD files are created using the same RRD definition from gmetad.conf. All metrics are automatically aggregated to the cluster and grid views.
In practice, an administrator may want more control over this behavior. For each metric received, the administrator should be able to choose various options: a) ignore the metric b) choose from multiple possible RRD file formats or RRD file locations c) customise the path and filename d) not include the metric in the aggregation process. A practical use case may involve storing some metrics on disk while storing others on ramdisk, or storing short-interval values (every 20 seconds) in ramdisk and storing hourly and daily min/max/avg values to file. This may require using libconfuse configuration parser for the gmetad.conf file to represent all these possibilities. This could also be implemented through the Python gmetad instead of the regular gmetad. -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org